Internal Dialogue During Divergent Thinking is Associated With Originality

Abstract

Internal dialogue (ID) may play a key self-regulatory function in creative thinking. This study extends previous research by contrasting the contributions of trait-level ID—reflecting general tendencies—and state-level ID, reported retrospectively after completing divergent thinking (DT) tasks. Three hundred fifty adults completed two DT tasks, which were scored for fluency and originality. After accounting for individual differences in openness to experience (OE), trait-level ID was inconsistently related to originality. In contrast, state-level ID was consistently related to originality, even after controlling for trait-level ID and OE. Contrary to our hypothesis, however, this association was negative. Higher levels of ID during DT were related to lower originality. This unexpected pattern suggests that ID may not always support original ideation. Individual differences in ID were not associated with fluency, suggesting that its potential role applies more to idea quality than to quantity. Alternative explanations for these findings, limitations, along with discrepancies with prior research, are examined to guide future studies into the potential role of ID in creative thinking.

Keywords

creativity divergent thinking inner speech internal dialogue openness to experience

Over the past several decades, creativity has attracted interest and gained popularity across diverse disciplines (Lubart et al., 2013). It is considered a uniquely human characteristic, as well as a valuable asset for solving everyday problems and effectively navigating professional contexts (Thornhill-Miller et al., 2023). Given its relevance, it is not surprising that researchers are focused on estimating creative potential, typically assessed through performance on divergent thinking tasks (Barbot et al., 2016; Runco, 2008; Runco & Acar, 2012)—measures that capture one's ability to generate numerous and original responses to open-ended prompts (Guilford, 1967). Studying the mechanisms underlying this potential is an important research objective and has prompted significant interest in identifying the individual differences that shape performance on divergent thinking tasks. In the present study, we extend this line of research by further exploring internal dialogue as a key individual difference. Prior research suggests that trait-level differences—that is, individuals’ daily experiences of internal dialogue—are associated with performance on divergent thinking tasks (de Rooij, 2023b). In this study, we propose that state-level internal dialogue, or the self-reported extent of dialogue occurring during the task itself, may serve as a more sensitive indicator of originality.

In 2022, the Organisation for Economic Co-operation and Development (OECD) has included the assessment of creative thinking in its Program for International Student Assessment (PISA). Published in 2024, the PISA report defines creative thinking as the ability to generate, evaluate, and improve original and varied ideas (OECD, 2024). It is most often evaluated through divergent thinking assessments. The Alternate Uses Test (AUT; Guilford, 1967) is by far the most widely used task paradigm, featuring in nearly 50% of divergent thinking studies (Saretzki et al., 2026). This task requires participants to generate novel uses for everyday objects within a limited time. Responses are typically evaluated in terms of fluency (i.e., number of responses) and originality. However, given that originality is a core component in the broader consensus definition of creativity, it has become central to the assessment of divergent thinking (Weiss et al., 2021) and is the primary measure of creative potential in most contemporary studies (Saretzki et al., 2026). Originality encompasses the ingenuity, rarity or uniqueness of the ideas produced in response to a given prompt (Reiter-Palmon et al., 2019; Saretzki et al., 2026). While such task-based assessments certainly capture key aspects of creative thinking, a deeper understanding also demands exploring how individual differences shape and constrain creative potential.

According to the multivariate approach to creativity, each individual possesses a unique profile shaped by the interplay of cognitive, conative, personality, and environmental factors (Barbot et al., 2016; Lubart et al., 2015). These components interact to meet the specific demands of divergent thinking tasks (and other creative tasks) across diverse contexts, thereby giving rise to an individual's creative potential. This potential is conceptualized as the convergence of multiple, distinct yet interconnected resources, closely related to key individual differences. The psychology of creativity has long emphasized the role of individual differences as contributors to creative potential. Such differences encompass a broad spectrum of personal characteristics, reflecting the unique ways through which individuals engage in the creative process. For instance, openness to experience—characterized by a broad appreciation for art, imagination, curiosity, novelty, and diverse experiences—is the personality trait most consistently and robustly associated with creativity and divergent thinking (Christensen, 2020; McCrae, 1987; Silvia et al., 2008). It has been proposed that individual differences in openness to experience tap into variability in self-regulation tendencies, with individuals higher in openness being more likely to use cognitive reappraisal strategies in their activities (Gresham & Gullone, 2012). This emphasizes how individual differences may explain, at least in theory, how people engage creatively.

Critically, there is general agreement that creative thinking involves the self-regulation of thoughts and actions, as supported by evidence from both experimental (e.g., Zielińska et al., 2025) and individual-differences perspectives (e.g., Dygert & Jarosz, 2020). One way these self-regulatory processes may manifest is through inner speech (Rieber & Carton, 1987) and internal dialogue (Valsiner, 2002). While the two concepts are strongly related, they differ in important ways. Individuals may experience inner voices (for a more detailed discussion, see Lind, 2025), whereas internal dialogue reflects more specifically the intrapersonal communication among these inner voices (Hermans, 1996).

Specifically, the concepts of inner speech and internal dialogue can be traced back to Vygotskian theory (Rieber & Carton, 1987), which proposes that adults verbally engage with children from an early age and thereby serve as external agents of self-regulation. Over time, children transform these social interactions into overt private speech, that is, speech directed toward themselves aloud, and eventually into inner speech through a process of internalization. Inner speech is thus generally defined as silent self-directed talk, in contrast to overt private speech, which is externally audible (Rieber & Carton, 1987). Like overt private speech, inner speech can take different forms and serve a variety of functions, including self-criticism, self-reinforcement, self-management, and social evaluation (Brinthaupt et al., 2009). While private speech refers to overt self-directed language, its progressive internalization gives rise to inner speech. According to Fernyhough's dialogic model (1996), inner speech can be monologic, but it may also retain the dialogic structure of social interactions, thereby giving rise to internal dialogues in which multiple internalized perspectives, or inner voices, are simulated. It is also distinct from imagined interactions, a broader construct that refers to a form of daydreaming in which individuals experience conversations accompanied by verbal and nonverbal features (Honeycutt, 2020; Morin, 2020). In the current study, we focus specifically on internal dialogue, defined as verbal intrapersonal communication among inner voices (Hermans, 1996).

Internal dialogue is not a unitary phenomenon. As described by Olés and Puchalska-Wasyl (2011, p. 242), individuals may engage in dialogues with imagined figures, simulate social dialogical relationships in their thoughts, and confront different points of view corresponding to various “I-positions” relevant to their personal and social identities. All forms of internal dialogue appear to serve a self-regulatory function, although the specific forms individuals characteristically engage in may influence how they think and behave in the world (Oleś, 2009; Oleś & Puchalska-Wasyl, 2010). While many individuals tend to experience inner speech in a dialogical form (Alderson-Day et al., 2018; McCarthy-Jones & Fernyhough, 2011), this tendency varies substantially across individuals. In particular, internal dialogue may be associated with personality traits. For instance, Puchalska-Wasyl et al. (2008) reported that dialogical inner activity reflects fundamental dimensions of personality and is moderately correlated with openness to experience (r = .44). In other words, individuals who frequently engage in internal dialogues also tend to report higher levels of self-awareness, fantasy, and aesthetic appreciation than those whose inner speech is primarily monological.

Building on the potential self-regulatory function of inner speech and dialogue, several recent lines of research from de Rooij have begun to explore its covariation with creative thinking (de Rooij, 2022, 2023a, 2023b; de Rooij et al., 2024). More specifically, results suggest that inner speech is involved in how individuals process uncertainty during idea generation tasks. In one such study, participants were asked to generate ideas on how a railway company could increase ridership. Inner speech appeared to support self-reinforcing thoughts under uncertainty, potentially triggering internal dialogue that in turn influenced originality (de Rooij, 2023a). Additional evidence from an idea evaluation task showed that inner speech was linked to confidence while participants assessed the creativity of solutions for attracting new customers (de Rooij, 2024). However, these studies did not focus specifically on divergent thinking, thereby limiting the inferences that can be drawn about the role of inner speech and internal dialogue in conventional measures of creative thinking.

Focusing more directly on divergent thinking, de Rooij (2022) investigated inner speech using the AUT, with idea originality assessed via automated scoring of semantic distance. Inner speech was measured using the Varieties of Inner Speech Questionnaire – Revised (Alderson-Day et al., 2018; McCarthy-Jones & Fernyhough, 2011), a 26-item self-report instrument assessing multiple dimensions of inner speech but including five items related to its dialogical nature. The study found no significant association between any aspect of inner speech—including dialogical activity—and the semantic distance of responses on the AUT. In a more recent study, however, de Rooij (2023b) adopted a more direct approach to internal dialogue by using the Internal Dialogical Activity Scale (IDAS; Oleś, 2009), which was used to estimate the intensity of internal dialogue individuals experience on a daily basis. Divergent thinking was assessed using two AUT prompts, with originality scored via a uniqueness-based technique. Regression analyses revealed no significant association between the extent of reported daily internal dialogue and the quantity of ideas generated during the divergent thinking tasks. However, the analysis did reveal a significant positive association between overall internal dialogical activity and idea originality, indicating that the intensity of trait-level internal dialogue is related to idea quality rather than mere quantity. Specifically, none of the dialogical subtypes predicted originality independently, suggesting that their effects were driven by shared variance in trait-level dialogical activity. The same study also examined the association between internal dialogue and “eminent” creative achievements (CAQ; Carson et al., 2005). Interestingly, trait-level internal dialogue correlated more strongly with originality in the AUT than with CAQ scores, reinforcing its potential as an indicator of creative potential rather than broader creative achievements.

While the results reported by de Rooij (2023b) offer encouraging support for the association between internal dialogue and creative ideation, it is important to note that the observed correlation between trait-level internal dialogue and originality was relatively modest (r = .16). In the present study, we propose that examining internal dialogue at the state level—by asking individuals to retrospectively assess the extent of their dialogical activity during a specific creative task—may show a stronger association with originality than trait-level measures. One commonly cited argument linking dialogical activity to enhanced original thinking is that it helps individuals maintain self-regulatory engagement and manage uncertainty during idea generation (de Rooij, 2023b). Alternatively, internal dialogue has been proposed to facilitate perspective-taking and the strategic deployment of cognitive resources (Fernyhough, 2016, pp. 108–109; Glaveanu, 2017 in de Rooij, 2022), which in turn may benefit divergent thinking. From this perspective, reliance on internal dialogue can be seen as a form of regulatory process activated during creative ideation—something that may not be fully captured by trait-level assessments. In other words, there may be a gap between how frequently individuals report engaging in internal dialogue in daily life and the extent to which they engage in it during specific moments of creative tasks. However, these regulatory processes are inherently difficult to observe directly. Determining what a person is thinking during ideation—and how those thoughts influence idea generation—remains a methodological challenge.

Gilhooly et al. (2007) sought to address this challenge by using a think-aloud paradigm to capture the cognitive strategies employed during divergent thinking tasks. By instructing participants to verbalize their thoughts while generating new uses for familiar objects, the authors were able to externally access aspects of participants’ unfolding mental processes. Although their methodology did not explicitly target internal dialogue, it represents a valuable step toward a more precise understanding of the role of self-directed thought in creativity. Building on a similar approach, Berthiaume et al. (2025) asked children from 3rd to 5th grade to self-report, after completing divergent thinking tasks, how they had generated each of their ideas, showing the feasibility of retrospectively asking individuals to reflect on their prior thought processes, even at young ages. These studies suggest that it might be feasible to assess internal dialogue during the execution of creative tasks—at least retrospectively—as a more immediate indicator of the regulatory mechanisms underlying originality.

Present Study

The present study is a partial replication of the work by de Rooij (2023b). While de Rooij relied on trait-level measure of internal dialogue, revealing significant yet modest associations with divergent thinking, we propose a methodological adaptation by instead using a state-level assessment of internal dialogue, that is, as it occurs during the completion of a creative task. If internal dialogue indeed contributes to the regulatory processes that support the generation of original ideas, then a state-level measure may better account for variation in creative thinking than a trait-level measure, which reflects broader and more stable individual tendencies. With this in mind, we developed a short state-level internal dialogue questionnaire, comprising four items adapted from the IDAS (Oleś, 2009). This instrument was designed to capture experiences and intensity of internal dialogue in direct relation to the ideation process, as elicited in the specific context of a divergent thinking task. We provide additional context regarding the development of this short scale in the Methods section.

The present study pursued two primary objectives. First, we aimed to partially replicate the findings of de Rooij (2023b) by examining the association between trait-level internal dialogue and the originality of ideas generated during a divergent thinking task. Second, we sought to evaluate whether state-level internal dialogue—assessed retrospectively after the task—is more strongly associated with originality than the trait-level measure. This approach aligns with efforts to develop a more contextualized assessment of internal dialogue by capturing the extent of dialogical activity as it occurs in response to a specific creative task. Moreover, given that openness to experience has been consistently linked to both creativity (McCrae, 1987; Silvia et al., 2009) and internal dialogue (Puchalska-Wasyl et al., 2008), we included it in our study as a control variable. In line with this, we predicted that average state-level internal dialogue would show a significant positive correlation with originality in divergent thinking tasks, independent of trait-level internal dialogue and openness to experience. In addition, we similarly examined the association between internal dialogue and fluency to further support the idea that internal dialogue is more closely related to idea quality (i.e., originality) than to quantity (i.e., fluency), consistent with previous research (de Rooij, 2023b).

Method

Participants

A total of 350 individuals between the ages of 18 and 65 years (M = 35.10, SD = 10.88) participated in the study, and included 167 women, 176 men, and 7 non-binary individuals. In terms of education levels, 71.5% of participants held a university degree, 16% had attended university without completing a degree, 10.2% had completed high school or a lower level of education, and 2.3% held a professional degree. Participants were recruited via Prolific, an online platform for crowdsourced recruitment that is designed for academic research and maintains strict internal control procedures for data quality. Although online data has been heavily criticized (Webb & Tangney, 2024), it is worth noting that the quality of data on Prolific has proven superior to other comparative platforms such as mTurk (Peer et al., 2022). Eligibility criteria required participants to be at least 18 years old and residents of the United States.

We estimated the required sample size using a power analysis conducted in GPower (version 3.1.9.7). The analysis was based on a minimal expected effect size (f² = 0.026), corresponding to a prior study that reported a correlation of r = .16 between internal dialogue and originality (de Rooij, 2023b). Assuming an alpha level of .05 and statistical power of .80, results indicated that a sample of 304 participants would be needed to detect the incremental variance explained by state-level internal dialogue, above and beyond that explained by trait-level dialogue and openness to experience.

Materials

Divergent Thinking

This was assessed using the (AUT; Guilford, 1967), a widely used measure of divergent thinking. Participants were presented with a prompt featuring a common object and asked to generate as many novel and original uses for it as possible, following standard be-creative instructions (Nusbaum et al., 2014). Following prior recommendations (Saretzki et al., 2026; Silvia et al., 2008) each prompt was administered with a two-minute time limit. Participants completed two prompts, randomly selected from the following list: (1) trophy, (2) coin, (3) sponge, and (4) bandage. To ensure data quality in the divergent thinking tasks, a research assistant reviewed all responses for invalid entries. Of the 3,349 entries assessed, 241 (7%) were excluded. Upon closer examination, approximately half of these excluded entries were either incomplete or included participant comments indicating idea fatigue (e.g., “I don’t have ideas anymore”). When these cases are removed, the proportion of genuinely invalid or off-task responses is reduced to approximately 3.5% of the total dataset. Responses were aggregated across both prompts for analysis. In total, participants generated 3,108 valid responses, with an average of 5.31 ideas per participant (SD = 2.04, range = 1–10).

In line with prior research examining the association between internal dialogue and divergent thinking (de Rooij, 2023b), we employed a uniqueness-like scoring method to assess originality. Each idea generated by participants was coded to allow a frequency-based analysis. Two trained research assistants independently reviewed all responses and assigned standardized labels to qualitatively similar ideas, ensuring that identical or equivalent responses were grouped under a single label. This approach facilitated the evaluation of idea frequency, thereby allowing us to quantify the uniqueness of each response—defined as the inverse of its frequency within the dataset. To ensure unbiased coding, all identifying variables related to other study measures (e.g., trait and state variables, openness to experience) were excluded from the dataset. The uniqueness score for each idea was computed as the reciprocal of its frequency (i.e., 1 / idea frequency). Accordingly, a unique idea—produced by only one participant in the sample of 350 individuals—was assigned a score of 1.0, whereas an idea mentioned twice received a score of 0.5 (i.e., 1 / 2 = 0.5). The most frequently reported ideas received lower scores, close to zero. This method minimizes confounding with fluency while preserving sensitivity to the distributional rarity of responses (for a review of assessment methods, see Reiter-Palmon et al., 2019). An average uniqueness score was then computed for each participant by averaging the uniqueness values of ideas across both prompts. From now on, this measure will be referred to as the frequency-based originality score.

To enhance the robustness of originality assessment in this study, we incorporated a second scoring method alongside the frequency-based originality score. This method used Ocsai* (version 1.6, based on gpt-4o), an open-access platform leveraging large language models to automatically estimate the originality of responses (Organisciak et al., 2023). This system is designed to replicate human-rated responses based on datasets containing several thousand examples. There is initial evidence that Ocsai provides valid scores, with ai-based scores correlating highly with human ratings (r = .81). Recent psychometric investigations into the validity of automated scoring systems for divergent thinking have shown that these approaches provide comparable or slightly higher levels of construct validity evidence than human ratings (Saretzki & Benedek, 2026). This system has been shown to be scalable across a range of creative tasks and prompts (Acar et al., 2024), languages (Zielińska et al., 2023), and age groups (e.g., de Chantal et al., 2026). In this study, the Ocsai-based originality scores were aggregated across both prompts.

Trait-Level Internal Dialogue

We assessed internal dialogical activity using an adapted version of the Revised Internal Dialogical Activity Scale (IDAS-R; Oleś et al., 2025), a self-report measure designed to capture the frequency of everyday internal dialogues. Participants rated a series of statements on a 5-point scale ranging from 1 (never) to 5 (very often). The IDAS-R was developed following a confirmatory factor analysis (CFA) of the original IDAS, resulting in a revised 40-item version (Oleś et al., 2025). A shorter 24-item version, the IDAS-R24, was derived by selecting items with the highest factor loadings. The IDAS-R24 has demonstrated solid psychometric properties, including good construct validity and reliability, reported as superior to the 40-item version. Cronbach's alpha values indicate acceptable to high internal consistency across subscales, and test-retest reliability (RTT), assessed over a two-week interval, supports the temporal stability of the measure. The eight IDAS-R24 subscales and their reliability indices are as follows: Identity Dialogues (α = .87, RTT = .82), Maladaptive Dialogues (α = .72, RTT = .65), Social Dialogues (α = .83, RTT = .80), Supportive Dialogues (α = .84, RTT = .80), Spontaneous Dialogues (α = .85, RTT = .82), Ruminative Dialogues (α = .82, RTT = .60), Confronting Dialogues (α = .81, RTT = .72), and Change of Perspective (α = .79, RTT = .72).

Given the time constraints of the present study, we employed a reduced 12-item version derived from the IDAS-R24. This decision was guided by our focus on overall involvement in internal dialogical activity rather than on specific subtypes, as is more common in psychotherapeutic or clinical assessment contexts (Oleś et al., 2025). Our aim was therefore to assess participants’ general propensity to engage in internal dialogue, rather than to test hypotheses concerning distinct dialogical forms. This strategy was further supported by prior findings indicating that no individual IDAS subtype showed a stronger association with originality in divergent thinking tasks than the others (de Rooij, 2023b). Accordingly, we selected the four subscales with the strongest factor loadings in the revised measure—Identity Dialogues, Spontaneous Dialogues, Confronting Dialogues, and Change of Perspective. Eleven of the retained items had factor loadings above .68 in the original scale, whereas one item, despite a lower loading (.49), was preserved because of its relevance to the present study. More specifically, these selected subscales were those accounting for the largest proportion of variance in the original instrument (Oleś et al., 2025). By contrast, Maladaptive, Social, Supportive, and Ruminative Dialogues were not included in the shortened version.

Although this reduction necessarily narrows the coverage of dialogical subtypes, it remains aligned with the present study's focus on general dialogical involvement and, more importantly, with self-regulatory accounts of the role of internal dialogue in creative thinking. In particular, Identity Dialogues, Spontaneous Dialogues, Confronting Dialogues, and Change of Perspective appear especially relevant in the context of divergent thinking, as it has been proposed to support perspective taking, self-reflection, and the strategic deployment of cognitive resources (Fernyhough, 2016, pp. 108–109; Glaveanu, 2017 in de Rooij, 2022). In addition, previous work has highlighted the role of creative self-identity in the expression of creative potential (Karwowski et al., 2018), as well as the importance of confronting and evaluating one's own ideas (Guo et al., 2022). Taken together, these methodological and conceptual considerations indicate that the selected subtypes likely capture the aspects of internal dialogue most relevant to the present research question.

Consistent with this rationale, responses were aggregated into a single composite score reflecting overall internal dialogical activity. The full set of selected items and their factor loadings is reported in Appendix B. In the present sample, the 12-item version showed high internal consistency, α = .88, 95% CI [.86, .89], indicating that the abbreviated measure retained strong reliability despite its reduced length. For transparency, we also report pairwise correlations between each retained subtype and the other study variables (see Appendix; Table A1). As shown in Table A1, all four retained subtypes were associated with fluency and both originality indices in the same direction, with only minor differences in statistical significance. In other words, no retained subtype showed a pattern of association with divergent thinking that was meaningfully distinct from that of the others. Taken together, these converging results suggest that, although the shortened version does not reproduce the full multidimensional structure of the original scale, it may preserve the general fluctuation in trait-level internal dialogical activity targeted in the present study.

State-Level Internal Dialogue

State-level internal dialogue was assessed using a brief 4-item questionnaire presented on a 5-point scale ranging from 1 (never) to 5 (very often). The items were adapted from the IDAS-R24 (Oleś et al., 2025) to reflect participants’ possible experiences specifically during the divergent thinking task (see Appendix C). This adapted instrument was designed to capture internal dialogical activity from a situational, task-specific perspective, providing an overall measure of dialogical intensity. As with the trait-level measure, our aim was to assess general internal dialogical activity rather than any defined subtype. To develop this state-level instrument, we examined whether items from each IDAS-R24 subscale could be meaningfully reformulated to reflect experiences occurring during task performance. This process indicated that items from two subscales—Spontaneous Dialogues and Change of Perspective—could be translated most clearly into a state-level format. The selected items were therefore chosen for their contextual relevance and fit with the divergent thinking task and were reworded to refer explicitly to participants’ experiences during the two preceding AUT prompts.

For example, the original statement “I converse with myself” was adapted as “During the creative task, I engaged in a dialogue with myself.” This item was intended to capture a general tendency to engage in dialogical activity during task performance. Similarly, the statement “When I have a difficult choice, I talk the decision over with myself from different points of view” was reformulated as “During the creative task, when faced with a difficult choice, I engaged in self-talk and considered it from multiple perspectives.” Although the items were derived from the Spontaneous Dialogues and Change of Perspective subscales, subtype scores were not analyzed separately in the present study. As a consequence, although informed by the items from the trait-level IDAS-R24, the state-level measure constitutes a distinct, task-specific operationalization of internal dialogical activity.

In the current study, the four adapted items showed satisfactory internal consistency (α = .76, 95% CI [.72, .80]). For transparency, we also report pairwise correlations between each retained subtype and the other study variables (see Appendix A). As shown in Table A1, the two subtypes were associated with fluency and both originality scores quite similarly. In other words, no subtype showed an association with divergent thinking that was meaningfully different from the associations observed for the other subtype. These results support our rationale for combining the two subtypes into a single composite score of state-level internal dialogue.

Openness to Experience

The HEXACO Personality Inventory (Ashton & Lee, 2011) is a psychometric instrument developed to assess six core personality dimensions, including openness to experience. Sixteen items targeting this trait, with half reversed-coded, were selected and presented on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). In the current study, the items showed excellent internal consistency (α = .85, 95% CI [.830, .875]).

Procedure

Participants were recruited online through the Prolific platform, and the study was administered via Qualtrics. The study began with two AUTs, with prompts selected randomly from a pool of four objects. Immediately afterward, participants completed the state-level internal dialogue questionnaire, which assessed their dialogical activity experienced during the creative task. This was followed by the trait-level internal dialogue questionnaire, the openness to experience measure, and a brief sociodemographic questionnaire. The entire procedure took approximately 10 min to complete.

Analytical Strategy

Consistent with our working hypothesis, the mean score of the state-level internal dialogue (ID)—reflecting participants’ self-reported engagement in inner dialogue during the divergent thinking task—was expected to show a stronger association with ideational originality than trait-level ID and openness to experience (OE). The dependent variable in this analysis was the mean originality score across both divergent thinking tasks, operationalized using both frequency- and Ocsai-based scoring methods. Although the results from analyses using both originality scores generally converged, we chose to present them side by side to enhance the robustness and transparency of our findings. We also conducted an identical analysis using mean fluency scores. The independent variables included state-level ID, trait-level ID, and OE, with the latter serving primarily as a control variable. As previously mentioned, OE is recognized within the Big Five model as the personality trait most consistently associated to creativity (Silvia et al., 2009), and is also known to correlate with internal dialogical activity (de Rooij, 2023b). Zero-order correlations between all study variables, along with corresponding descriptive statistics, are presented in Table 1. Additional correlations involving specific subtypes of the trait- and state-level ID measures, although not examined in detail in the present study, are reported in the Appendix (see Table A1).

Table 1.
Zero-order Correlations Between Study Variables.

M (SD) 1 2 3 4 5 6

1. Trait-level ID 3.27 (0.72) — [.48, .63] [.03, .24] [−.21, −.01] [−.21, .00] [−.22, −.01]

2. State-level ID 3.22 (0.88) .56*** — [−.03, .18] [−.21, .00] [−.28, −.08] [−.25, −.05]

3. OE 3.66 (0.60) .14* .08 — [−.11, .10] [−.16, .05] [.02, .23]

4. Fluency 5.31 (2.04) −.11* −.10 −.01 — [.18, .38] [−.03, .18]

5. Freq-based originality 0.21 (0.14) −.11* −.18* −.06 .28 — [.35, .52]

6. Ocsai-based originality 2.50 (0.62) −.11* −.15** .13* .08 .44*** —

Notes. ID: Internal dialogue. OE: Openness to experience. Values above the diagonal represent 95% confidence intervals.

* p < .05. p < .01. * p < .01.

Statistical analyses were performed using IBM SPSS Statistics, version 29. To examine the contribution of state-level ID to originality, we conducted a hierarchical regression analysis. This approach allowed us to incrementally isolate the variance explained by state-level ID beyond other relevant variables. Prior to conducting the regression analyses, we verified the assumptions required for linear regression. All assumptions were satisfactorily met. We implemented two hierarchical linear regression models, each designed to address a specific research objective.

Model 1 aimed to partially replicate findings from de Rooij (2023b) by examining the predictive contribution of trait-level ID and OE on frequency- and Ocsai-based originality, as well as on fluency. Model 2 questioned whether state-level ID is a stronger predictor of originality and fluency than the trait-level measure. In this model, state-level ID was thus added as an independent variable, while trait-level ID and OE remained as controls. This approach allowed us to isolate the respective contributions of state- and trait-level measures of ID. Critically, the model tested whether state-level ID explained a significant portion of variance in originality and fluency scores beyond that already accounted for by the trait-level measure of ID and OE.

Results

Analysis of Zero-Order Correlations

Analysis of Table 1 revealed several noteworthy correlations among the study variables. Most relevant to the current study, trait- and state-level internal dialogue (ID) were found to be moderately correlated (r = .56, p < .001), supporting the idea that these measures assess related psychological constructs. Openness to experience (OE) was positively associated with trait-level ID, further justifying its role as a control variable in the regression models. Additionally, there was a moderate correlation between frequency- and Ocsai-based originality scores (r = .44, p < .001), providing convergent support for the relative reliability of these alternative scoring methods.

Fluency was negatively correlated with trait-level internal dialogue but showed no statistically significant association with state-level ID. Both originality measures were significantly associated with the composite scores of trait- and state-level ID. Correlations with specific ID subtypes are reported in the Appendix (see Table A4). Although one trait-level subtype was not significantly associated with originality, all other subtypes—at both trait and state levels—showed consistent associations with originality measures, all in the same (negative) direction. Subsequent analyses will provide a more robust test of the predictive power of ID.

Is Trait-Level Internal Dialogue Associated with Originality and Fluency?

Frequency-based originality as dependent variable. Model 1 accounted for 1.4% of the variance in the frequency-based originality score. OE did not significantly predict originality (β = −.05, B = −.01, 95% CI −.13, .05], t = −0.85, p = .396), and after controlling for OE, trait-level ID was also not a significant predictor (β = −.10, B = −.02, 95% CI [−.20, .01], t = −1.91, p = .057). However, it should be noted that trait-level ID emerged as a significant predictor when OE was excluded from the model, suggesting that OE may act as a confounding variable in the association between trait-level ID and originality. Additional model statistics are reported in Table A2.

Ocsai-based originality as dependent variable. When the dependent variable was the Ocsai-based originality score, the model explained 3.3% of the variance, reaching statistical significance, F(2, 347) = 5.95, p < .01. In this case, both OE (β = .14, B = .15, 95% CI [.04, .25], t = 2.70, p < .01) and trait-level ID (β = −.13, B = −.11, 95% CI [−.23, −.03], t = −2.49, p = .013) were significant, independently predicting changes in originality. Unexpectedly, a one standard deviation (SD) increase in trait-level ID was associated with a 0.13 SD decrease—not the expected increase—in Ocsai-based originality, even after controlling for OE. Additional model statistics are reported in Table A3.

Fluency as dependent variable. When the dependent variable was the fluency indicator, the model explained 1.2% of the variance, but did not reach significance, F(2, 347) = 2.21, p = .11. Additional model statistics are reported in Table A4.

Is State-Level Internal Dialogue Associated with Originality and Fluency Beyond Trait-Level Differences?

Frequency-based originality as dependent variable. In Model 2, state-level ID was added as a predictor to assess its unique contribution to frequency-based originality. This model accounted for 3.4% of the variance in originality scores, representing a significant improvement over the first model, ΔR² = 2.0%, F(1, 346) = 7.21, p < .01. As in Model 1, neither OE (β = −.05, B = −.01, 95% CI [−.03, .01], t = −0.85, p = .40) nor trait-level ID (β = −.01, B = −.01, 95% CI [−.03, .02], t = −0.11, p = .91) significantly predicted originality. However, state-level ID emerged as a significant negative predictor (β = −.17, B = −.03, 95% CI [−.05, −.01], t = −2.68, p = .008). Specifically, a one standard deviation increase in state-level ID was associated with a 0.17 SD decrease in originality, reflecting a small to medium effect size. Additional model statistics are reported in Table A2.

Ocsai-based originality as dependent variable. When the dependent variable was the Ocsai-based originality score, a similar pattern of results emerged. The full model explained 4.5% of the variance and reached statistical significance, ΔR² = 1.2%, F(1, 346) = 4.24, p = .04. In this model, trait-level ID was not a significant predictor of originality (β = −.06, B = −.05, 95% CI [−.16, .06], t = −0.93, p = .351). OE, however, remained a significant and independent predictor (β = .14, B = .15, 95% CI [.04, .25], t = 2.72, p < .01). Importantly, after controlling for both trait-level ID and OE, state-level ID also emerged as a significant predictor of Ocsai-based originality (β = −.13, B = −.10, 95% CI [−.18, −.01], t = –2.06, p = .04), again reflecting a small-to-medium negative effect size. Additional model statistics are reported in Table A3.

Fluency as dependent variable. When fluency was used as the dependent variable, the full regression model explained only 1.5% of the variance and was not statistically significant, ΔR² = 0.2%, F(1, 346) = 0.81, p = .37. Within this model, none of the predictors—trait-level ID, OE or state-level ID—significantly predicted fluency. Additional model statistics are reported in Table A4.

Discussion

There is growing interest in the role that internal dialogue (ID) may play in creative ideation. Previous research has shown that trait-level ID—defined as the extent to which individuals experience dialogical activity in daily life—is significantly associated with originality in divergent thinking tasks (de Rooij, 2023b), with no specific subtype showing stronger association than the others. However, the strength of this association has been modest. In the present study, we sought to replicate and extend these findings by examining a state-level measure of ID—specifically, the intensity to which individuals engaged in dialogical activity during divergent thinking tasks. We hypothesized that the state-level measure would show a stronger association with originality, and that it would account for unique variance in originality beyond the contributions of trait-level ID and openness to experience—a personality trait known to be associated with both internal dialogue (Puchalska-Wasyl et al., 2008) and divergent thinking (Silvia et al., 2008). To provide a more robust test of our hypothesis, we employed two distinct yet correlated methods for assessing originality. One method used recent artificial intelligence algorithms (Organisciak et al., 2023), while another analyzed the uniqueness of the ideas, aligning more closely with the scoring paradigm used by de Rooij (2023b). Additionally, we included an analysis of fluency to further disentangle the contribution of internal dialogue to idea quality (originality) as opposed to idea quantity.

Divergent Thinking is More Strongly Related to State- Than Trait-Level Internal Dialogue

In our sample of 350 individuals, we found that the zero-order correlations between trait-level internal dialogue and originality were consistent across both originality scoring methods but of relatively small magnitude. These correlations were weaker than those reported by de Rooij (2023b), whose estimates fell near the upper bound of the absolute confidence interval observed in the present study. However, after controlling for openness to experience, trait-level ID was significantly associated with originality in divergent thinking only when originality was operationalized using Ocsai-based scores. This suggests that the contribution of trait-level ID to originality is relatively unstable and may depend on the method used to assess originality, potentially accounting for some of the inconsistencies observed in the existing literature (de Rooij, 2022, 2023b). The results related to state-level ID were notably more consistent. State-level ID was significantly associated with differences in originality beyond trait-level ID and openness to experience, regardless of the originality scoring method employed. In addition, neither trait- nor state-level internal dialogue was associated with fluency in the divergent thinking tasks after controlling for openness to experience, further reinforcing that internal dialogue is more strongly related to idea quality than to idea quantity.

But More Internal Dialogue is Associated with Lower Originality

In a certain way, our hypothesis was supported—state-level ID showed a stronger association with originality than the combined contribution of trait-level ID and openness to experience. However, an unexpected finding emerged. Contrary to our initial hypothesis, the association between dialogical activity and divergent thinking was negative. Individuals who reported higher levels of ID during the task tended to produce less original responses. As such, we failed to replicate the findings of de Rooij (2023b), who reported a positive, rather than negative, association between trait-level ID and originality. Our findings contrast with those reported by de Rooij (2023b), and unexpected results should therefore be interpreted with caution. In the following, we outline several possible interpretations of these findings.

Such divergence may be due to methodological differences between the two studies. One possible factor involves the characteristics of the divergent thinking tasks. Although both studies employed the same task paradigm—i.e., AUTs with standard “be creative” instructions—the specific objects presented differed. In fact, prior research has shown that variations in objects can influence originality (e.g., Beaty et al., 2022), suggesting that task prompts could introduce variability in creative ideation. In the same way, both studies employed uniqueness-based scoring methods to assess originality, although the specific implementation details differed. However, while such differences may contribute to variations in effect sizes, they are unlikely to explain the reversal in the direction of the association between ID and originality.

Another potential explanation lies in differences in how ID activity was assessed. In the present study, we used a shortened version of the IDAS, selecting items not only based on their factor loadings, but also in light of a conceptual rationale emphasizing the self-regulatory functions of internal dialogue in divergent thinking. From this perspective, the selected subscales appeared especially relevant, given, for example, prior arguments concerning the role of creative self-identity in the expression of creative potential (Karwowski et al., 2018) and the importance of evaluating and confronting one's own ideas during creative thinking (Guo et al., 2022). Although this adaptation reduced the number of items, the resulting scale showed high internal consistency. We fully acknowledge, however, that shortening the instrument may have altered the breadth of the construct captured relative to the full scale.

At the same time, several observations suggest that item reduction alone is unlikely to account for the core discrepancy in findings. First, the shortened trait-level ID scale retained the expected positive association with Openness to Experience, consistent with prior research (Puchalska-Wasyl et al., 2008), and it was also positively correlated with the state-level measure, despite only partial overlapping in ID subtypes. Second, we did not observe any substantial differences in the patterns of association of the specific subtypes with either fluency or originality. Rather, the results were broadly consistent with de Rooij's (2023b) findings in suggesting that no single subtype appears to relate to divergent thinking in a meaningfully distinct way relative to the others. Although it remains possible that including all subtypes would have produced a somewhat different overall pattern, particularly in terms of effect size, it seems less likely that this would have reversed the direction of the observed associations. A more cautious interpretation, therefore, is that originality in divergent thinking was negatively associated with the forms of dialogical activity captured by the subtypes retained in the present study. Accordingly, further research is needed to clarify under which conditions, and for which forms of internal dialogue, dialogical activity may be positively, negatively, or not reliably associated with divergent thinking.

One additional explanation for the divergent results may lie in procedural differences, particularly in the sequencing of study components. In de Rooij's (2023b) study, participants completed the trait-level ID questionnaire before engaging in the divergent thinking tasks. In contrast, in the current study, participants completed both the state- and trait-level ID measures after performing the AUT. It is conceivable—though still speculative—that completing the trait-level questionnaire beforehand may have activated participants’ awareness of their internal dialogue, thereby influencing the degree or type of cognitive engagement during the task. This priming effect could have prompted more deliberate or structured use of dialogical activity, potentially enhancing its observable association with originality. A similar priming effect may also have occurred within the current study. Specifically, because the internal dialogue questionnaires were administered immediately after the divergent thinking tasks, prior engagement in creative thinking may have influenced participants’ subsequent reports of their internal dialogue. Although such explanations do not fully account for the patterns observed in either the state- or trait-level internal dialogue measures, they offer a possible framework within which the findings can be tentatively interpreted, and which could certainly guide future research.

Directions for Future Research

Beyond procedural explanations, it is possible that participants with lower originality scores reported higher levels of ID due to greater difficulty generating creative ideas. Indeed, ID is often activated in response to uncertainty, particularly when individuals face ambiguity, complexity, or a lack of clear knowledge (Hermans & Hermans-Konopka, 2010). Divergent thinking tasks require sustained cognitive effort to produce responses that are both novel and appropriate (de Chantal et al., 2025; Nusbaum & Silvia, 2011). Participants struggling to generate satisfactory or original ideas may be more prone to self-questioning and evaluative internal discourse. In this context, ID may function as a regulatory process aimed at reducing uncertainty through intrapersonal reflection. Therefore, heightened ID during the task may reflect reactive self-monitoring and critical evaluation, especially when participants perceive their ideas as insufficiently original. It is possible that prompting participants to reflect on their internal dialogue prior to engaging in a divergent thinking task prepares them to use internal dialogue proactively, rather than reactively in response to the uncertainty and difficulty of the task. We believe that future research should aim to address these critical nuances.

Finally, another interesting possibility, suggested by a reviewer's remark, is that internal dialogue during idea generation may be related to a greater disposition toward mind wandering, defined as the occurrence of task-unrelated thoughts (Mooneyham & Schooler, 2013). Importantly, although some studies suggest that mind wandering may facilitate creativity during incubation periods, when individuals temporarily disengage from the task (Sio & Ormerod, 2009), other studies indicate that it may be detrimental to originality during idea generation itself (e.g., Hao et al., 2015). While research on the association between internal dialogue and mind wandering remains limited, at least to our knowledge, one recent study found that greater inner speech was associated with greater mind wandering (Racy & Morin, 2024). Our rationale and core hypothesis were grounded in the idea that the self-regulatory function of internal dialogue would support divergent thinking. At the same time, it is possible that individuals reporting higher levels of state-level internal dialogue are also more prone to task-unrelated thoughts during the task. This possibility clearly warrants further investigation into the associations among internal dialogue, mind wandering, and creative thinking.

Strengths and Limitations

The present study has several notable strengths and limitations. The sample size of 350 participants, balanced across gender, provided sufficient statistical power to detect small to moderate effect sizes, enhancing the reliability of the findings. The administration of the divergent thinking tasks followed best practices to enhance reliability and validity. Participants completed two randomly selected prompts from a pool of four, and each task was preceded by be-creative instructions, which is key to maximizing the validity of divergent thinking tasks. The convergence of our results across two distinct, yet related, scoring methods further reinforces the robustness and reliability of the findings. Finally, by including a validated measure of openness to experience, the study was able to control for a key personality variable known to be associated with both ID and creativity, thereby strengthening the interpretability of the findings.

Finally, a key limitation concerns the inherent difficulty of measuring inner experience. Internal dialogue falls within this domain and is, by its nature, not always easy to identify or articulate with precision. This issue may be especially relevant in the present study, as participants were asked to report on their internal dialogue retrospectively rather than through experience-sampling methods, which are often used to capture inner experience closer to the moment of occurrence. Moreover, the very act of attending to and describing inner experience is often unfamiliar to participants (Hurlburt & Heavey, 2001). As with any unfamiliar task, repeated practice or training may be necessary to support more reliable and accurate introspective reports. Accordingly, it should be acknowledged that participants likely differed in their ability to recognize and report their internal dialogue accurately.

Conclusion

Creativity is increasingly recognized as a critical skill and has become a central focus within psychological research. Creative potential is typically assessed through a combination of cognitive and dispositional factors, including divergent thinking and other individual differences. In this study, we examined the association between internal dialogue and divergent thinking, extending previous research that reported a positive correlation between the two (de Rooij, 2023b). Our findings were unexpected. State-level internal dialogue—assessed immediately after the divergent thinking task—showed a stronger association with originality than trait-level internal dialogue, which reflects a general daily tendency. However, this association was negative, with participants reporting more internal dialogue during the task producing ideas of lower originality. This result suggests a more complex and context-dependent association between internal dialogue and creativity, potentially influenced by procedural factors and the nature of cognitive engagement during ideation. Despite being unexpected, these findings contribute to the growing body of literature on the cognitive and conative dimensions underlying creativity. They suggest the need for further investigation into the role played by internal dialogue in creative thinking and call for more nuanced approaches to its study.

	M (SD)	1	2	3	4	5	6
1. Trait-level ID	3.27 (0.72)	—	[.48, .63]	[.03, .24]	[−.21, −.01]	[−.21, .00]	[−.22, −.01]
2. State-level ID	3.22 (0.88)	.56***	—	[−.03, .18]	[−.21, .00]	[−.28, −.08]	[−.25, −.05]
3. OE	3.66 (0.60)	.14*	.08	—	[−.11, .10]	[−.16, .05]	[.02, .23]
4. Fluency	5.31 (2.04)	−.11*	−.10	−.01	—	[.18, .38]	[−.03, .18]
5. Freq-based originality	0.21 (0.14)	−.11*	−.18***	−.06	.28**	—	[.35, .52]
6. Ocsai-based originality	2.50 (0.62)	−.11*	−.15**	.13*	.08	.44***	—

Footnotes

Acknowledgments

We would like to thank Madeleine Gourd and Sarah Haidar for their assistance in scoring the divergent thinking tasks. We would also like to thank Alwin de Rooij and Oleś Piotr for sharing their study materials.

ORCID iDs

Christian Labrie

Pier-Luc de Chantal

Ethics Approval Statement

This study was approved by the Human Research Ethics Committee of the Université du Québec à Montréal [2022-4231].

Funding

This work was supported by a grant from the Social Sciences and Humanities Research Council of Canada [430-2022-00931] and the Fonds de recherche du Québec [2024-2025-NP-345487] to PLDC.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Data and research materials are freely available on the Open Science Framework ().

Appendix A

Appendix B

Appendix C

References

Acar

Dumas

Organisciak

Berthiaume

(2024). Measuring original thinking in elementary school: Development and validation of a computational psychometric approach. Journal of Educational Psychology, 116(6), 953–981. https://doi.org/10.1037/edu0000844

Alderson-Day

Mitrenga

Wilkinson

McCarthy-Jones

Fernyhough

(2018). The varieties of inner speech questionnaire – Revised (VISQ-R): Replicating and refining links between inner speech and psychopathology. Consciousness and Cognition, 65, 48–58. https://doi.org/10.1016/j.concog.2018.07.001

Ashton

Lee

(2011). HEXACO Personality inventory-revised [Dataset]. https://doi.org/10.1037/t03138-000

Barbot

Besançon

Lubart

(2016). The generality-specificity of creativity: Exploring the structure of creative potential with EPoC. Learning and Individual Differences, 52, 178–187. https://doi.org/10.1016/j.lindif.2016.06.005

Beaty

R. E.

Johnson

D. R.

Zeitlen

D. C.

Forthmann

(2022). Semantic distance and the alternate uses task: Recommendations for reliable automated assessment of originality. Creativity Research Journal, 34(3), 245–260. https://doi.org/10.1080/10400419.2022.2025720

Berthiaume

K. C.

Acar

Dumas

(2025). “My brain said that … ”: A qualitative study of sources of children’s creative ideas. Creativity Research Journal, 37(4), 687–711. https://doi.org/10.1080/10400419.2024.2336739

Brinthaupt

T. M.

Hein

M. B.

Kramer

T. E.

(2009). The self-talk scale: Development, factor analysis, and validation. Journal of Personality Assessment, 91(1), 82–92. https://doi.org/10.1080/00223890802484498

Carson

S. H.

Peterson

J. B.

Higgins

D. M.

(2005). Reliability, validity, and factor structure of the Creative Achievement Questionnaire. Creativity Research Journal, 17(1), 37–50. https://doi.org/10.1207/s15326934crj1701_4

Christensen

A. P.

(2020). Openness to experience. In The Palgrave encyclopedia of the possible (pp. 1–7). Springer International Publishing. https://doi.org/10.1007/978-3-319-98390-5_113-1

10.

de Chantal

P.-L.

Leblanc

Houde-Labrecque

Labrie

(2026). Cognitive control and creativity in children: The role of task-specific demands in divergent thinking. The Journal of Creative Behavior, 60(1), e70095. https://doi.org/10.1002/jocb.70095

11.

de Chantal

P.-L.

Zielińska

Lebuda

Karwowski

(2025). How do examples impact divergent thinking? The interplay between associative and executive processes. Psychology of Aesthetics, Creativity, and the Arts, 19(5), 1148–1163. https://doi.org/10.1037/aca0000613

12.

de Rooij

(2022). Varieties of inner speech and creative potential. Imagination, Cognition and Personality, 41(4), 460–489. https://doi.org/10.1177/02762366211070999

13.

de Rooij

(2023a). Inner speaking and uncertainty during idea generation. The Journal of Creative Behavior, 57(3), 376–396. https://doi.org/10.1002/jocb.584

14.

de Rooij

(2023b). Internal dialogue, creative potential, and creative achievement. Imagination, Cognition and Personality, 43(2), 105–128. https://doi.org/10.1177/02762366231173608

15.

de Rooij, A. (2024). Speaking to your inner muse: How self-regulation by inner speaking influences confidence during idea evaluation. Creativity Research Journal, 36(2), 291-308. https://doi.org/10.1080/10400419.2022.2124356

16.

de Rooij

Atef

Faber

(2024). A wandering mind is not always a creative mind: How thought dynamics explain the relationship between mind wandering and creativity. The Journal of Creative Behavior, 58(1), 151–170. https://doi.org/10.1002/jocb.640

17.

Dygert

S. K. C.

Jarosz

A. F.

(2020). Individual differences in creative cognition. Journal of Experimental Psychology: General, 149(7), 1249–1274. https://doi.org/10.1037/xge0000713

18.

Fernyhough

(1996). The dialogic mind: A dialogic approach to the higher mental functions. New Ideas in Psychology, 14(1), 47–62. https://doi.org/10.1016/0732-118X(95)00024-B

19.

Fernyhough

(2016). The voices within: The history and science of how we talk to ourselves. Basic Books.

20.

Gilhooly

K. J.

Fioratou

Anthony

S. H.

Wynn

(2007). Divergent thinking: Strategies and executive involvement in generating novel uses for familiar objects. British Journal of Psychology, 98(4), 611–625. https://doi.org/10.1111/j.2044-8295.2007.tb00467.x

21.

Glaveanu, V. (2017). The creative self in dialogue. In M. Karwowski & J. C. Kaufman (Eds.), The creative self (pp. 117-135). Academic Press.

22.

Gresham

Gullone

(2012). Emotion regulation strategy use in children and adolescents: The explanatory roles of personality and attachment. Personality and Individual Differences, 52(5), 616–621. https://doi.org/10.1016/j.paid.2011.12.016

23.

Guilford

J. P.

(1967). The nature of human intelligence. McGraw-Hill.

24.

Guo

Lin

Acar

Jin

Feng

Zeng

(2022). Divergent thinking and evaluative skill: A meta-analysis. The Journal of Creative Behavior, 56(3), 432–448. https://doi.org/10.1002/jocb.539

25.

Hao

Runco

M. A.

Pina

(2015). More mind wandering, fewer original ideas: Be not distracted during creative idea generation. Acta Psychologica, 161, 110–116. https://doi.org/10.1016/j.actpsy.2015.09.001

26.

Hermans

Hermans-Konopka

(2010). Dialogical self theory: Positioning and counter-positioning in a globalizing society (1st ed.). Cambridge University Press. https://doi.org/10.1017/CBO9780511712142

27.

Hermans

H. J. M.

(1996). Voicing the self: From information processing to dialogical interchange. Psychological Bulletin, 119(1), 31–50. https://doi.org/10.1037/0033-2909.119.1.31

28.

Honeycutt

J. M.

(2020). Imagined interactions and inner speech. Imagination, Cognition and Personality, 39(4), 386–396. https://doi.org/10.1177/0276236619869163

29.

Hurlburt

R. T.

Heavey

C. L.

(2001). Telling what we know: Describing inner experience. Trends in Cognitive Sciences, 5(9), 400–403. https://doi.org/10.1016/S1364-6613(00)01724-1

30.

Karwowski

Lebuda

Wiśniewska

(2018). Measuring creative self-efficacy and creative personal identity. The International Journal of Creativity & Problem Solving, 28(1), 45–57.

31.

Lind

(2025). Are there really people with No inner voice? Commentary on nedergaard and Lupyan (2024). Psychological Science, 36(9), 765–767. https://doi.org/10.1177/09567976251335583

32.

Lubart

Mouchiroud

Tordjman

Zenasni

(2015). Psychologie de la créativité—2e édition: 2e éd. Armand Colin. Cairn.info. https://doi.org/10.3917/arco.lubar.2015.01

33.

Lubart

Zenasni

Barbot

(2013). Creative potential and its measurement. International Journal for Talent Development and Creativity, 1(2), 41–51.

34.

McCarthy-Jones

Fernyhough

(2011). The varieties of inner speech: Links between quality of inner speech and psychopathological variables in a sample of young adults. Consciousness and Cognition, 20(4), 1586–1593. https://doi.org/10.1016/j.concog.2011.08.005

35.

McCrae

R. R.

(1987). Creativity, divergent thinking, and openness to experience. Journal of Personality and Social Psychology, 52(6), 1258–1265. https://doi.org/10.1037/0022-3514.52.6.1258

36.

Mooneyham

B. W.

Schooler

J. W.

(2013). The costs and benefits of mind-wandering: A review. Canadian Journal of Experimental Psychology / Revue Canadienne de Psychologie Expérimentale, 67(1), 11–18. https://doi.org/10.1037/a0031569

37.

Morin

(2020). When inner speech and imagined interactions meet. Imagination, Cognition and Personality, 39(4), 374–385. https://doi.org/10.1177/0276236619864276

38.

Nusbaum

E. C.

Silvia

P. J.

(2011). Are intelligence and creativity really so different? Fluid intelligence, executive processes, and strategy use in divergent thinking. Intelligence, 39(1), 36–45. https://doi.org/10.1016/j.intell.2010.11.002

39.

Nusbaum

E. C.

Silvia

P. J.

Beaty

R. E.

(2014). Ready, set, create: What instructing people to “be creative” reveals about the meaning and mechanisms of divergent thinking. Psychology of Aesthetics, Creativity, and the Arts, 8(4), 423–432. https://doi.org/10.1037/a0036549

40.

OECD. (2024). PISA 2022 Results (Volume III): Creative minds, creative schools. OECD. https://doi.org/10.1787/765ee8c2-en

41.

Oleś

P. K.

(2009). Czy głosy umysłu da się mierzyć? Skala Wewnętrznej Aktywności Dialogowej (SWAD) [Are voices of the mind measurable? Internal dialogical activity scale (IDAS)]. Prz. Psychology, 52, 37–50.

42.

Oleś

P. K.

Kutnik

Marczewska

Pasternak

(2025). Internal dialogical activity and how to measure it: Internal dialogical activity scale-revised. Journal of Constructivist Psychology, 38(4), 600–615. https://doi.org/10.1080/10720537.2024.2446756

43.

Oleś

P. K.

Puchalska-Wasyl

(2010). Between nature and culture: Dialogicality as a basic human feature (pp. 41–51).

44.

Oleś

P. K.

Puchalska-Wasyl

(2011). Dialogicality and personality traits. In Hermans

H. J. M.

Gieser

(Eds.), Handbook of dialogical self theory (pp. 241–252). Cambridge University Press. Cambridge Core. https://doi.org/10.1017/CBO9781139030434.017

45.

Organisciak

Acar

Dumas

Berthiaume

(2023). Beyond semantic distance: Automated scoring of divergent thinking greatly improves with large language models. Thinking Skills and Creativity, 49, 101356. https://doi.org/10.1016/j.tsc.2023.101356

46.

Peer

Rothschild

Gordon

Evernden

Damer

(2022). Data quality of platforms and panels for online behavioral research. Behavior Research Methods, 54(4), 1643–1662. https://doi.org/10.3758/s13428-021-01694-3

47.

Puchalska-Wasyl

Chmielnicka-Kuter

Oleś

(2008). From internal interlocutors to psychological functions of dialogical activity. Journal of Constructivist Psychology, 21(3), 239–269. https://doi.org/10.1080/10720530802071476

48.

Racy

Morin

(2024). Relationships between self-talk, inner speech, mind wandering, mindfulness, self-concept clarity, and self-regulation in university students. Behavioral Sciences, 14(1), 55. https://doi.org/10.3390/bs14010055

49.

Reiter-Palmon

Forthmann

Barbot

(2019). Scoring divergent thinking tests: A review and systematic framework. Psychology of Aesthetics, Creativity, and the Arts, 13(2), 144–152. https://doi.org/10.1037/aca0000227

50.

Rieber

R. W.

Carton

A. S.

(Eds.). (1987). The collected works of L. S. Vygotsky: Problems of general psychology, including the volume thinking and speech. Springer US. https://doi.org/10.1007/978-1-4613-1655-8

51.

Runco

M. A.

(2008). Commentary: Divergent thinking is not synonymous with creativity. Psychology of Aesthetics, Creativity, and the Arts, 2(2), 93–96. https://doi.org/10.1037/1931-3896.2.2.93

52.

Runco

M. A.

Acar

(2012). Divergent thinking as an indicator of creative potential. Creativity Research Journal, 24(1), 66–75. https://doi.org/10.1080/10400419.2012.652929

53.

Saretzki

Benedek

(2026). Investigating the validity evidence of automated scoring methods for divergent thinking assessments. Creativity Research Journal, 1–17. https://doi.org/10.1080/10400419.2025.2607669

54.

Saretzki

Forthmann

Benedek

(2026). A systematic quantitative review of divergent thinking assessments. Psychology of Aesthetics, Creativity, and the Arts, 20(1), 1–20. https://doi.org/10.1037/aca0000691

55.

Silvia

P. J.

Nusbaum

E. C.

Berg

Martin

O’Connor

(2009). Openness to experience, plasticity, and creativity: Exploring lower-order, high-order, and interactive effects. Journal of Research in Personality, 43(6), 1087–1090. https://doi.org/10.1016/j.jrp.2009.04.015

56.

Silvia

P. J.

Winterstein

B. P.

Willse

J. T.

Barona

C. M.

Cram

J. T.

Hess

K. I.

Martinez

J. L.

Richard

C. A.

(2008). Assessing creativity with divergent thinking tasks: Exploring the reliability and validity of new subjective scoring methods. Psychology of Aesthetics, Creativity, and the Arts, 2(2), 68–85. https://doi.org/10.1037/1931-3896.2.2.68

57.

Sio

U. N.

Ormerod

T. C.

(2009). Does incubation enhance problem solving? A meta-analytic review. Psychological Bulletin, 135(1), 94–120. https://doi.org/10.1037/a0014212

58.

Thornhill-Miller

Camarda

Mercier

Burkhardt

J.-M.

Morisseau

Bourgeois-Bougrine

Vinchon

El Hayek

Augereau-Landais

Mourey

Feybesse

Sundquist

Lubart

(2023). Creativity, critical thinking, communication, and collaboration: Assessment, certification, and promotion of 21st century skills for the future of work and education. Journal of Intelligence, 11(3), 54. https://doi.org/10.3390/jintelligence11030054

59.

Valsiner

(2002). Forms of dialogical relations and semiotic autoregulation within the self. Theory & Psychology, 12(2), 251–265. https://doi.org/10.1177/0959354302012002633

60.

Webb

M. A.

Tangney

J. P.

(2024). Too good to be true: Bots and bad data from Mechanical Turk. Perspectives on Psychological Science, 19(6), 887–890. https://doi.org/10.1177/17456916221120027

61.

Weiss

Steger

Kaur

Hildebrandt

Schroeders

Wilhelm

(2021). On the trail of creativity: Dimensionality of divergent thinking and its relation with cognitive abilities, personality, and insight. European Journal of Personality, 35(3), 291–314. https://doi.org/10.1002/per.2288

62.

Zielińska

Lebuda

Czerwonka

Karwowski

(2025). Self-regulation prompts improve creative performance. The Journal of Creative Behavior, 59(1), e674. https://doi.org/10.1002/jocb.674

63.

Zielińska

Organisciak

Dumas

Karwowski

(2023). Lost in translation? Not for Large Language Models: Automated divergent thinking scoring performance translates to non-English contexts. Thinking Skills and Creativity, 50, 101414. https://doi.org/10.1016/j.tsc.2023.101414

Internal Dialogue During Divergent Thinking is Associated With Originality—But in Unexpected Ways

Abstract

Keywords

Present Study

Method

Participants

Materials

Divergent Thinking

Trait-Level Internal Dialogue

State-Level Internal Dialogue

Openness to Experience

Procedure

Analytical Strategy

Results

Analysis of Zero-Order Correlations

Is Trait-Level Internal Dialogue Associated with Originality and Fluency?

Is State-Level Internal Dialogue Associated with Originality and Fluency Beyond Trait-Level Differences?

Discussion

Divergent Thinking is More Strongly Related to State- Than Trait-Level Internal Dialogue

But More Internal Dialogue is Associated with Lower Originality

Directions for Future Research

Strengths and Limitations

Conclusion

Footnotes

Acknowledgments

ORCID iDs

Ethics Approval Statement

Funding

Declaration of Conflicting Interests

Data Availability Statement

Appendix A

Appendix B

Appendix C

References