*Result*: Generative insight: Aha! Moments in category generation and divergent thinking.

*Do the aha ! moments of insight observed in creative problem solving also occur during divergent thinking and other generative tasks? We hypothesized that when participants think of novel and creative ideas, those moments may be accompanied by the same subjective feelings that describe insight experiences. Participants generated lists of responses for each of 12 cues that included taxonomic categories (e.g., metals ), ad hoc categories (e.g., things that are green ), and divergent thinking prompts (e.g., alternative uses of bricks ), clicking a red button labeled " aha! " whenever they had ideas that were surprising, sudden, and unexpected. Results showed that insight moments were commonly reported during divergent thinking and that aha ! ideas were more creative and novel than non- aha ! ideas and that aha ! moments were observed in all generative tasks, both creative and noncreative. Furthermore, aha ! ideas, as compared with non- aha ! responses, were preceded by much longer pauses, with the duration of these nascent pauses increasing during the course of idea generation. These findings indicate that the cognitive processes involved in nascent pauses and insight moments should be incorporated into neurocognitive models of ideation. (PsycInfo Database Record (c) 2026 APA, all rights reserved).*

*

Generative Insight: Aha! Moments in Category Generation and Divergent Thinking / BRIEF REPORT

<cn> <bold>By: Steven M. Smith</bold>
> Department of Psychological and Brain Sciences, Texas A&M University
> <bold>Visheeta J. Chandolia</bold>
> Department of Psychological and Brain Sciences, Texas A&M University
> <bold>Matthew A. Kidd</bold>
> Department of Psychological and Brain Sciences, Texas A&M University
> <bold>Morgan S. Paladino</bold>
> Department of Psychological and Brain Sciences, Texas A&M University </cn>

<bold>Acknowledgement: </bold>Kimberly Fenn served as action editor.The data and ideas of this article have been presented at the annual Association for Research in Memory, Attention, Decision-making, Imagery, Learning, Language, and Organization Conference at Texas State University in October 2023 and at the annual meeting of the Psychonomic Society in November 2023 in San Francisco, California. The authors declare that they have no competing interests. This project was funded by the Department of Psychological and Brain Sciences at Texas A&M University.Steven M. Smith played a lead role in conceptualization, investigation, methodology, project administration, resources, supervision, writing–original draft, and writing–review and editing. Visheeta J. Chandolia played a lead role in formal analysis and a supporting role in data curation, investigation, methodology, software, and writing–review and editing. Matthew A. Kidd played a supporting role in data curation, formal analysis, and writing–review and editing. Morgan S. Paladino played a supporting role in data curation, formal analysis, and writing–review and editing.

Many studies of creative ideation can be sorted into two paradigms: (a) the study of divergent production, that is, the ability to fluently and flexibly produce a set of creative ideas related to a cue (e.g., Beaty & Silvia, 2012; Benedek & Neubauer, 2013; Madore et al., 2015; Zabelina & Robinson, 2010) and (b) the study of insight, that is, the sudden and unexpected awareness of the solution to a problem (e.g., Duncker, 1945; Maier, 1931; Metcalfe, 1986). A dichotomous view of creative cognition (e.g., Dechaume et al., 2024; Eysenck, 2003; Hommel et al., 2011; Kim & Pierce, 2013; Runco & Acar, 2012; Shen et al., 2018) distinguishes convergent thinking, where problem solving produces a single correct solution (e.g., Remote Associates Test or insight problems), from divergent thinking, where many diverse, unusual, and creative ideas are encouraged (e.g., Alternative Uses Task). The flaw in the common dichotomous view is the process-pure assumption, the assumption that each task is purely a test of a single cognitive process (e.g., Jacoby, 1991). In a similar discussion about conflating tasks (direct and indirect tests of memory) with cognitive processes (intentional and automatic uses of memory), Jacoby (1991) noted that “problems interpreting task dissociations have arisen from equating particular processes with particular tasks and then treating those tasks as if they provide pure measures of those processes” (p. 513). Along with other researchers (e.g., Eymann et al., 2024), we challenge the strict distinction between insight problem solving and divergent production, and we propose that bringing creative ideas to mind (the goal of divergent thinking) and bringing creative solutions of problems to mind (the goal of insight problem solving) share important qualities.

The experience of insight in creative problem solving is a sudden, rapid, and unexpected awareness of the solution to a problem. Hallmarks of aha! moments include an initial experience of fixation while inappropriate cognitive structures are explored, a protracted pause prior to the insight, and rapid restructuring when the solution appears because the problem-solver sees the problem differently. Restructuring might involve, for example, going beyond the ordinary function of an object and using it in a novel way (e.g., Maier, 1931), or recognizing and relaxing a counterproductive implicit constraint (e.g., Knoblich et al., 2001). In the context of generative ideation, restructuring might involve, for example, shifting from a conventional retrieval plan for generating four-footed animals (e.g., listing four-footed pets or farm animals) to a novel retrieval plan (e.g., listing four-footed dinosaurs or cartoon characters). Thus, generative restructuring might involve shifting to an unexpected retrieval plan after conventional plans have stopped producing new ideas. This version of generative restructuring is consistent with semantic clustering and foraging models of generative ideation (e.g., Benedek et al., 2023; Nijstad et al., 2010; Tellez et al., 2024); it predicts aha! moments when shifts between clusters of responses occur, particularly after conventional categories have been exhausted.

A feature of aha! experiences is a pause prior to announcing an insight. For example, Schooler et al. (1993) found that verbalization during problem solving, using a think-aloud protocol (e.g., Ericsson & Simon, 1980), benefited analytic problem solving but impaired insight problem solving; Schooler and Melcher (1995) further reported that participants often stopped verbalizing prior to announcing insight solutions and that such pauses were greater for insight problems than for analytic ones. Salvi and Bowden (2024) attributed such discontinuities to switching from conscious to unconscious thought in the time leading up to an aha! moment. Similarly, Baird et al. (2012) described such discontinuities as a shift from a state of focused attention to a mind-wandering state, and Jacobs and Metcalfe (2024) described a shift from seeking knowledge in deliberate ways that engage a reinforcement-based system (Curiosity-1) to a more open-ended pursuit of knowledge (Curiosity-2). We refer to the time during this hypothesized shifted state that precedes insight moments as a nascent period, meaning simply the time when an insight is emerging. A nascent pause, that is, a response gap just before an insight report is given, would likely be filled with fixation on the previous retrieval cluster, followed by one of the shifts described above that would enable the sudden nondeliberate retrieval of an unexpected novel idea. If insight experiences are preceded by discontinuities in cognition and if generation of insightful solutions in creative problem solving resembles generation of creative ideas in divergent thinking, then pauses between responses in generative thinking should be longer preceding aha! responses than non-aha! responses.

Three different types of generation tasks were used: listing members of taxonomic categories (e.g., metals, countries), ad hoc categories (e.g., liquids, green things), and divergent thinking (e.g., uses for bricks or cardboard boxes). Because taxonomic categories are already known, representations of those categories were expected to be stable. Ad hoc categories and divergent thinking categories, which are created on demand rather than preexisting in semantic memory, are flexible and might be represented in more than one way (Chrysikou, 2006). The restructuring that might occur for ad hoc categories and divergent thinking tasks might result in insight moments, whereas we predicted that taxonomic categories would be less subject to restructuring and accompanying aha! moments. This prediction was also supported by findings of incubation effects for divergent thinking and ad hoc category generation tasks, but not taxonomic category generation (Smith et al., 2017).

We hypothesized that people experience insight or aha! moments when novel and creative ideas come to mind during divergent thinking and other generative tasks. Because a creative insight in problem solving might be experienced like getting a creative idea during divergent thinking (Chrysikou, 2006), we predicted that responses accompanied by aha! reports (aha! responses) would be statistically more novel, subjectively more creative, and preceded by longer pause times than non-aha! responses. In addition, we predicted that responses to ad hoc categories and divergent thinking cues would be more likely than responses to taxonomic categories to be reported as aha! moments.

Method

> <h31 id="xge-155-2-555-d69e232">Participants</h31>

Participants were undergraduates in introductory psychology courses earning partial credit for a course requirement; students self-enrolled on an online (Sona) website to participate, enrollment was voluntary, and all participants were briefed on other options for earning equal credit. Sample size was determined based on power analyses to obtain a moderate-to-large effect size (Cohen’s d = 0.80). A total of 64 students self-enrolled in the experiment. There were 12 participants who were excluded from analyses for either failing to respond to questions or not following instructions. The final sample consisted of 52 participants.

<h31 id="xge-155-2-555-d69e238">Design and Procedure</h31>

The experiment was run in a quiet laboratory room on campus with an experimenter present. Participants were consented and given a web link to begin the experiment. The experiment procedures, including instructions for the generation tasks, were done online using the PsychoPy platform. The entire experiment lasted approximately for 45 min.

Participants had a series of 12 generative tasks with 3 min for each task. The procedure began with four category generation tasks for taxonomic categories (four-footed animals, fruits, countries, metals), followed by four category generation tasks for ad hoc categories (liquids, things that make noise, things that are green, and things made of wood), and finally four divergent thinking tasks (alternative uses of bricks, alternative uses of tires, alternative uses of cardboard boxes, and alternative ways to improve their university).

The same order of tasks was used for all participants, with taxonomic category generation first, because we wanted to avoid expectancy effects for taxonomic cues. That is, if creative production from flexibly structured categories were necessary for aha! moments, we predicted we would not observe aha! moments in taxonomic category generation, especially because instructions for taxonomic category generation did not even mention creativity or creative ideas. Thinking of taxonomic category generation as our baseline or control condition, we did not want participants to be trying to think creatively in that condition. Thus, we did not mention “creative ideas” prior to taxonomic category generation, fearing to do so might encourage reports of creative insight. Therefore, taxonomic generation was always done first.

On the screen appeared the task prompt (e.g., “List all the four-footed animals you can think of”), a text window where responses were typed, and a large round red “button” with the word “aha!” in the center. Shown under the big red aha! button were the words, “Press AHA! when you have an unexpected idea.”

Instructions given to participants concerning aha! moments were the following: “In this experiment we are especially interested in what are called aha! moments. When people get ideas or solve problems, their ideas may be expected because they are planned or intended. On the other hand, sometimes ideas seem to ‘pop into mind’ or to ‘come out of nowhere,’ that is, they are surprising, sudden, and unexpected. Ideas that pop unexpectedly into mind are called aha! moments. We want to find out exactly when people experience these aha! moments. During the procedure, you will see a large red button labeled ‘AHA!’ Whenever you have an aha! moment, please click the red AHA! button as quickly as you can before you type the idea.”

Participants were also told, “For each prompt, type as many responses as you can during the 3 min you have. If you feel like you are running out of ideas, please keep trying for the entire 3 min, because more ideas will usually come to mind.”

Instructions for listing responses varied slightly for the three different types of generative tasks. For ad hoc categories and divergent thinking prompts, but not taxonomic cues, the instructions included the statement, “Include as many unusual and imaginative things as you can think of.”

<h31 id="xge-155-2-555-d69e275">Scoring</h31>

A frequency norm of responses for each writing cue was compiled for all 52 participants. Each response was given a frequency score depending upon the number of participants that gave that response. We calculated a novelty score for each response in the norm using the formula Novelty = [(N − response frequency) × 100]/(N − 1), where N = number of participants. Thus, a response given by every single participant would have a novelty score of 0, whereas a response given by only one participant would have a novelty score of 100. Two independent judges subjectively rated creativity on a scale of 1–5 for each response. Raters were asked to judge creativity of a response based on the degree to which the response referred to something not typically associated with the category (uncommon), if the response indicated an alternative but reasonable interpretation of the category (clever), and if they would have thought of the response as a participant (remote). Similar criteria have been used in the past for subjective creativity judgments (e.g., Christensen et al., 1957; Silvia et al., 2008) The raters’ scores showed good agreement (Intraclass Correlation Coefficient = 0.77).

<bold>Transparency and Openness</bold>

In the Method section, we report how we determined our sample size, all data exclusions, all manipulations, and all measures in the study. The complete data set and analysis code employed in the analyses of the presented study are available at <a href="https://doi.org/10.6084/m9.figshare.28232702.v2" target="_blank">https://doi.org/10.6084/m9.figshare.28232702.v2</a>. Research materials can be found in the Method section. The experiment was not preregistered.

Results

>

Every participant reported aha! moments during each of the 12 generation tasks, with an average of 2.47 aha! moments and 12.87 non-aha! responses for each 3-min task. Overall, 13% of participant responses were reported as aha! moments.

We compiled response norms for each of the 12 categories, including four taxonomic categories, four ad hoc categories, and four divergent thinking categories. Analyses included 2,346 different responses. Each participant generated an average of 20.09 responses for each of the four taxonomic category cues, 18.19 responses for each ad hoc category cue, and 8.53 responses for each divergent thinking cue.

Creativity and novelty scores of responses were positively correlated. For taxonomic categories, with a total of 417 different responses, the correlation was significant (r = .37, p &lt; .001). Correlations were also significant for ad hoc categories, with a total of 1,120 independent responses (r = .30. p &lt; .001), and for divergent thinking cues, with a total of 809 different responses (r = .22, p &lt; .001).

To analyze within-subject differences for a 3-min period of generation, participant responses were aggregated across conditions of response type (aha! or non-aha!) and cue type (taxonomic, ad hoc, or divergent thinking). Four 2 × 3 analysis of variances were then calculated to identify differences in four outcome measures (response frequency, pause time, novelty, and creativity). To analyze the time course of generation, four regression analyses were calculated to predict those measures, using the interval (i.e., first 30 s to sixth 30 s) and response type as predictors.

<h31 id="xge-155-2-555-d69e333">Frequency of Responses</h31>

There was a main effect of response type, F(1, 46) = 209.79, p &lt; .001, ηp<sups>2</sups> = 0.82; there were significantly more non-aha! responses (M = 13.11, SE = 0.46) than aha! responses (M = 3.59, SE = 0.46). There was a significant main effect of cue type, F(2, 98) = 126.04, p &lt; .001, ηp<sups>2</sups> = 0.72, with divergent thinking cues averaging the lowest total response frequency (M = 4.83, SE = 0.40), compared to ad hoc (M = 9.67, SE = 0.39); t(98) = −11.16, p &lt; .001, and taxonomic cues (M = 10.54, SE = 0.39); t(98) = −13.05, p &lt; .001. Finally, there was a significant interaction between response type and cue type, F(2, 83) = 67.53, p &lt; .001, ηp<sups>2</sups> = 0.62 (see Figure 1). Post hoc comparisons reveal the difference between aha! and non-aha! frequency had the smallest effect for divergent thinking, t(128) = 4.09, p = .001, d = 0.57; compared to ad hoc, t(128) = 12.92, p &lt; .001, d = 1.79; and taxonomic cues, t(128) = 16.01, p &lt; .001, d = 2.22.
>
><anchor name="fig1"></anchor>

A mixed-effects linear regression was calculated using interval of generation and its interaction with response type as fixed effects and included random effects for participants and cue type. Using Nakagawa et al.’s (2017) procedure on coefficients of determination, the fixed effects alone accounted for a considerable portion of variance (marginal R<sups>2</sups> = 0.39) and when including random effects (conditional R<sups>2</sups> = 0.59). The fixed effect of interval (b = −0.51, 95% CI [−0.53, −0.49]) was significant, indicating that response frequency decreased linearly throughout generation. Finally, the interaction between response type and 30-s interval was significant (b = 0.46, 95% CI [0.42, 0.50]), such that non-aha! frequency decreased at a greater rate than aha! frequency (Figure 2).
>
><anchor name="fig2"></anchor>

<h31 id="xge-155-2-555-d69e452">Pause Time Measure</h31>

Pause time for a given response was the time elapsed (in seconds) between submission of a response and submission of the previous response. There was a main effect of response type, F(1, 46) = 23.84, p &lt; .001, ηp<sups>2</sups> = 0.34, where aha! responses (M = 15.46, SE = 0.88) were preceded by significantly longer pauses than non-aha! responses (M = 9.06, SE = 0.88). There was a main effect of cue type, F(2, 98) = 48.81, p &lt; .001, ηp<sups>2</sups> = 0.49, with divergent thinking cues exhibiting the longest pauses (M = 17.51. SE = 0.82), compared to ad hoc (M = 10.06, SE = 0.81); t(98) = 7.76, p &lt; .001, and taxonomic cues (M = 9.21, SE = 0.81); t(98) = 8.57, p &lt; .001. Finally, the interaction between response type and cue type was not significant, F(2, 83) = 0.26, p = .774, ηp<sups>2</sups> = 0.01 (see Figure 3).
>
><anchor name="fig3"></anchor>

To analyze the effects of response type and interval on pause time, a mixed-effects linear regression was calculated and included random effects for participants and cue type. The fixed effects alone accounted for marginal R<sups>2</sups> = 0.127 of variance in pause time, while the full model (conditional R<sups>2</sups> = 0.279) accounted for about 30% of the variance. There was a significant effect for interval (b = 1.82, 95% CI [1.75, 1.89]), indicating that pause times increased linearly throughout generation. Finally, the interaction between response type and 30-s interval was significant (b = 1.65, 95% CI [1.48, 1.82]), where pauses increased at a greater rate for aha! responses throughout generation (Figure 4).
>
><anchor name="fig4"></anchor>

<h31 id="xge-155-2-555-d69e542">Novelty Measure</h31>

We found a significant main effect of response type, F(1,46) = 22.635, p &lt; .001, ηp<sups>2</sups> = .330, and cue type, F(2, 98) = 452.076, p &lt; .001, ηp<sups>2</sups> = .902, on novelty. Aha! responses (M = 76.10, SE = 1.01) were significantly more novel than non-aha! responses (M = 70.40, SE = 1.01). Responses to divergent thinking cues (M = 85.70, SE = 1.00) were significantly more novel, t(98) = 6.557, p &lt; .001, than responses to ad hoc cues (M = 78.40, SE = 0.98) and taxonomic cues (M = 55.70, SE = 0.99), t(98) = 26.694, p &lt; .001. Responses to ad hoc cues were also more novel than to taxonomic cues, t(98) = 20.924, p &lt; .001.The interaction between cue type and response type was not significant, F(2, 83) = 1.34, p = .27, ηp<sups>2</sups> = 0.031.

A linear mixed-effects regression showed a significant effect of response type (b = 8.140, 95% CI [5.005, 11.275], p &lt; .001) and interval (b = 5.618, 95% CI [5.272, 5.962], p &lt; .001) on novelty, such that aha! responses and responses in later intervals were more novel than non-aha! responses and early interval responses. The interaction between interval and response type was also significant (b = −2.302, 95% CI [–3.162, –1.442], p &lt; .001), suggesting that aha! responses increase less in novelty throughout generation than non-aha! responses. Overall, the model accounted for 36.7% variance in novelty (conditional R<sups>2</sups> = 0.367, marginal R<sups>2</sups> = 0.078; Table 1).
>
><anchor name="tbl1"></anchor>

<h31 id="xge-155-2-555-d69e647">Creativity Measure</h31>

There were significant main effects of response type, F(1, 46) = 13.010, p = .001, ηp<sups>2</sups> = 0.220, and cue type, F(2, 98) = 78.916, p &lt; .001, ηp<sups>2</sups> = 0.617, on creativity ratings. Aha! responses (M = 1.53, SE = 0.04) were rated as more creative than non-aha responses (M = 1.36, SE = 0.04). Responses to divergent cues (M = 1.790, SE = 0.044) were rated as more creative than those to ad hoc cues (M = 1.440, SE = 0.043), t(98) = 6.043, p &lt; .001. Ad hoc cues also produced higher creativity ratings than taxonomic cues (M = 1.100, SE = 0.043), t(98) = 6.324, p &lt; .001. Divergent thinking cues resulted in higher creativity ratings than taxonomic cues, t(98) = 12.095, p &lt; .001. The interaction between response type and cue type was not significant, F(2, 83) = 2.48, p = .09, ηp<sups>2</sups> = 0.056 (see Table 2).
>
><anchor name="tbl2"></anchor>

Discussion

>

The findings of this experiment include three discoveries: (a) Insight experiences occur during divergent thinking; (b) a nascent period, a very long pause (see Figure 3), precedes many aha! moments during divergent thinking; and (c) nascent pauses and insight moments occur in other generative ideation tasks, including taxonomic and ad hoc category generation.

The results demonstrate that aha! moments occur often during generative tasks such as divergent thinking and listing members of categories. Approximately one out of eight responses to taxonomic category cues were reported to be aha! moments, about one of five responses to ad hoc category cues were aha! moments, and one in four divergent thinking responses were aha! moments. This is a remarkably high frequency of insight experiences, given that heretofore such aha! moments have never been reported. The results show that divergent thinking tasks (e.g., the Alternative Uses Test) and convergent thinking tasks (e.g., solving Remote Associates Test problems) share an important element, the experience of insight.

Ideas linked to the aha! moments we observed were more novel and creative than ideas not accompanied by aha! phenomenology. Our instructions for listing responses to taxonomic category prompts did not include the goal of being “creative,” a goal that was included in our instructions for ad hoc category and divergent thinking prompts. Although we did not systematically examine the effects of such instructions on performance, it is clear that aha! moments are reported for responses from well-established categories, even without instructions to be creative.

The serial order effect, a robust and commonly observed phenomenon, shows that during divergent thinking the rate of responses decreases as time passes and that later ideas are relatively more novel and creative (Beaty & Silvia, 2012; Kohn & Smith, 2011). The present experiment found a serial order effect, that is, we found that the response rate decreased over the 3-min time course of generating responses for a cue and ideas generated at the end were more creative and novel than ideas given at the beginning of the interval. The present results go beyond what is known about the serial order effect; non-aha! responses decreased in frequency over time, but aha! responses did not decrease in frequency over the 3 min (Figure 2). This finding is consistent with the notion that people generate these two types of responses in different ways, resulting in different patterns of retrieval dynamics.

The time elapsed between successive responses, that is, the pause times or interresponse times, was considerably longer when they preceded aha! responses, as compared to non-aha! responses. We refer to this long pause time as the nascent period, the time during which aha! ideas are produced or discovered. These nascent periods grew longer at a faster rate than non-aha! pause times. A similar pattern of interresponse times occurs for recall of categorized lists of words; pauses between responses are longer when the two words are from different categories rather than the same category, and this difference grows longer as recall proceeds (e.g., Patterson et al., 1971; Rohrer & Wixted, 1994). Future research should investigate these similar patterns.

What happens during a nascent period that leads to an insight experience? We do not know, but there are several theories about a shift that may occur during such a period. An incremental theory (e.g., Roediger & Thorpe, 1978; Shiffrin, 1970) states that if responses in a set are randomly sampled (with replacement) via an iterative retrieval process, depletion of unsampled responses in the set over time could explain the serial order effect (fewer responses over time), and it would be consistent with our finding of increasing pauses between responses over time. A dual process theory, however, distinguishes deliberate retrieval from involuntary retrieval or “mind-popping” (e.g., Kvavilashvili & Mandler, 2004). Increasing failures to deliberately retrieve as-yet-unsampled responses may trigger a different type of retrieval process, one that is more likely to result in an unexpected aha! response. Of these two explanations, only the dual process account predicts longer pauses before aha! responses than non-aha! responses, because the nondeliberate retrieval process should be more likely than the deliberate one to result in unexpected responses. Such a shift from deliberate retrieval to being receptive to automatic retrieval might involve changing from a verbal mode to a nonverbal one (Schooler & Melcher, 1995), from a conscious state to a state driven by unconscious processes (Salvi & Bowden, 2024), from planful activity on a task to mind-wandering (Baird et al., 2012), from a reward-based system that drives one type of curiosity to a different type of curiosity (Jacobs & Metcalfe, 2024), from a brain state involving a cooperative relationship between the default mode network and the prefrontal executive control networks to a noncooperative state between those networks (Beaty et al., 2016), or from consideration of a problem using a fixated mindset to a new mindset in which interfering responses are rendered less accessible (Smith & Beda, 2024). There are also questions of what triggers a nascent pause and how a nascent period enables or produces the resultant insight.

Our findings indicate that insight moments, and the nascent pauses that precede many of them, occur not only for creative tasks but also during other generative tasks, including generation of category members from semantic memory. Theories of generative ideation, creative or otherwise (Benedek et al., 2023; Finke et al., 1992; Kounios & Beeman, 2009), have examined similarities and differences among various types of generation, including divergent thinking, problem solving, category generation, future simulation, and episodic memory recall; they should also account for the patterns of nascent pauses and insight moments that we have observed in generative ideation tasks. In the present study, we found remarkable similarity for all three generative tasks in terms of both the frequency of aha! ideas and the temporal dynamics for all three types of categories; the rate of non-aha! ideas decreased over the course of the 3-min generation task for all three tasks, and the difference in pause times between non-aha! and aha! responses increased duration of idea generation. Future research should examine the generality of unexpected content—aha! moments—in generative ideation and the temporal patterns of such content.

It is not difficult to understand why aha! moments occur in divergent thinking and ad hoc category generation tasks; responses can be ideas that are new to the participant and may be surprising because of their novelty. Responses generated from taxonomic categories, however, are drawn from preexisting semantic memories, not newly constructed ideas. Why would participants experience responses retrieved from semantic memory as “surprising, sudden, and unexpected,” as described in our instructions to participants? Our results cannot clearly answer this question, but we offer some speculative possibilities for aha! moments in taxonomic category generation. First, we note that our definition of aha! moments did not mention the need that responses be novel or creative. Why might retrieval of noncreative responses be experienced as sudden, surprising, and unexpected? One possibility is that a response might be drawn from an old or rare experience, a weak memory whose retrieval might be surprising. Another possibility is that a participant’s initial retrieval plan might change after a period of reflection, resulting in unexpected responses. Finally, there are studies of involuntary semantic memories, that is, known words or images that pop unexpectedly into mind without accompanying episodic contextual details (e.g., Kvavilashvili & Mandler, 2004). Further research is needed to discover why memories of members of taxonomic categories can be sudden, unintentional, and surprising, the hallmarks of aha! moments.

Could the aha! moments our participants reported be an artifact caused by expectancy effects that biased participants to give aha! reports only for their first idea after a long pause? If so, then aha! responses might not differ from non-aha! responses, aside from artifactual reports. Several facts argue against this possibility. First, the metacognitive study of insight and aha! moments has shown that subjective aha! reports correspond to more accurate solutions (e.g., Danek & Salvi, 2020; Salvi & Bowden, 2024; Salvi et al., 2016), better memory for solutions (Danek et al., 2013; Kizilirmak et al., 2016), and distinct brain states (e.g., Bowden & Jung-Beeman, 2003; Jung-Beeman et al., 2004). In the present study, aha! responses were more novel and creative than non-aha! responses. These findings run counter to the idea that insight and noninsight responses do not differ. In addition, in the present experiment, instructions to participants concerning aha! moments were similar to those used in insight problem-solving studies (Bowden & Jung-Beeman, 2003; Danek & Wiley, 2017; Danek & Salvi, 2020). Longer pauses have been reported prior to aha! solutions, relative to non-aha! solutions, in insight problem-solving studies (e.g., Salvi et al., 2015; Schooler & Melcher, 1995), findings consistent with those of the present experiment. Finally, as can be seen in Figure 4, pause times in the first 30-s interval were no different for aha! and non-aha! responses for taxonomic and ad hoc cues, and for divergent thinking cues, pause time was not diagnostic of aha! reports for the first three 30-s intervals. In addition, pause times for many non-aha! responses were quite long in later intervals. Therefore, the idea that long pauses trigger aha! reports artifactually is inconsistent with many of our results.

The finding that insight experiences occur often in divergent thinking contradicts a common distinction researchers make between convergent and divergent thinking measures, the former supposedly tested by insight problems and the latter by divergent thinking tests. A more unified theory of generative ideation describes both tasks as retrieving and constructing a series of responses, beginning with more fluent and accessible ideas, eventually leading to a nascent pause, and punctuated by an aha! moment in which a creative idea or solution comes to mind. Our results, and results of other studies (e.g., Eymann et al., 2022, 2024), undermine the common assumption that divergent thinking tasks, such as the Alternative Uses Test, and so-called “convergent thinking” tasks, such as insight problem solving, test qualitatively different types of creative thinking. Future research should abandon this flawed process-pure assumption and instead determine how to dissociate the roles of schema-driven or planned production of responses and non-schema-driven or unplanned discovery of ideas.

Finally, our findings may also be useful to insight researchers because our generation tasks evoke numerous aha! moments. Researchers of insight have benefited from the use of collections problems, such as Compound Remote Associates problems (Bowden & Jung-Beeman, 2003), insight problems (Metcalfe, 1986; Metcalfe & Wiebe, 1987), or matchstick arithmetic problems (Bilalić et al., 2021; Knoblich et al., 2001, 2005). Researchers usually find that participants solve several problems with insight, making insight an easy-to-produce phenomenon that can be studied experimentally. The present results show that generative paradigms can similarly produce frequent insight experiences; in our study more than one of every five divergent thinking responses was reported as an aha! moment.

<h31 id="xge-155-2-555-d69e971">Constraints on Generality</h31>

Our participants were college-age students enrolled in an introductory-level course fulfilling part of a course requirement by participating in our study. This population provided relative homogeneity and therefore less variance than a population more diverse in terms of age, educational level, and motivation. Although our conclusions are limited to this population, they nonetheless generalize to other studies of college-age participants, which are common.

References

<anchor name="c1"></anchor>

Baird, B., Smallwood, J., Mrazek, M. D., Kam, J. W., Franklin, M. S., & Schooler, J. W. (2012). Inspired by distraction: Mind wandering facilitates creative incubation. Psychological Science, 23(10), 1117–1122. 10.1177/0956797612446024

<anchor name="c52"></anchor>

Beaty, R. E., Benedek, M., Silvia, P. J., & Schacter, D. L. (2016). Creative cognition and brain network dynamics. Trends in Cognitive Sciences, 20(2), 87–95. 10.1016/j.tics.2015.10.004

<anchor name="c3"></anchor>

Beaty, R. E., & Silvia, P. J. (2012). Why do ideas get more creative across time? An executive interpretation of the serial order effect in divergent thinking tasks. Psychology of Aesthetics, Creativity, and the Arts, 6(4), 309–319. 10.1037/a0029171

<anchor name="c4"></anchor>

Benedek, M., Beaty, R. E., Schacter, D. L., & Kenett, Y. N. (2023). The role of memory in creative ideation. Nature Reviews Psychology, 2(4), 246–257. 10.1038/s44159-023-00158-z

<anchor name="c5"></anchor>

Benedek, M., & Neubauer, A. C. (2013). Revisiting Mednick’s model on creativity-related differences in associative hierarchies. Evidence for a common path to uncommon thought. The Journal of Creative Behavior, 47(4), 273–289. 10.1002/jocb.35

<anchor name="c6"></anchor>

Bilalić, M., Graf, M., Vaci, N., & Danek, A. H. (2021). The temporal dynamics of insight problem solving–restructuring might not always be sudden. Thinking & Reasoning, 27(1), 1–37. 10.1080/13546783.2019.1705912

<anchor name="c7"></anchor>

Bowden, E. M., & Jung-Beeman, M. (2003). Aha! Insight experience correlates with solution activation in the right hemisphere. Psychonomic Bulletin & Review, 10(3), 730–737. 10.3758/BF03196539

<anchor name="c54"></anchor>

Christensen, P. R., Guilford, J. P., & Wilson, R. C. (1957). Relations of creative responses to working time and instructions. Journal of Experimental Psychology, 53(2), 82–88. 10.1037/h0045461

<anchor name="c8"></anchor>

Chrysikou, E. G. (2006). When shoes become hammers: Goal-derived categorization training enhances problem-solving performance. Journal of Experimental Psychology: Learning, Memory, and Cognition, 32(4), 935–942. 10.1037/0278-7393.32.4.935

<anchor name="c9"></anchor>

Danek, A. H., Fraps, T., von Müller, A., Grothe, B., & Öllinger, M. (2013). Aha! experiences leave a mark: Facilitated recall of insight solutions. Psychological Research, 77(5), 659–669. 10.1007/s00426-012-0454-8

<anchor name="c10"></anchor>

Danek, A. H., & Salvi, C. (2020). Moment of truth: Why Aha! experiences are correct. The Journal of Creative Behavior, 54(2), 484–486. 10.1002/jocb.380

<anchor name="c53"></anchor>

Danek, A. H., & Wiley, J. (2017). What about false insights? Deconstructing the Aha! experience along its multiple dimensions for correct and incorrect solutions separately. Frontiers in Psychology, 7, 2077. 10.3389/fpsyg.2016.02077

<anchor name="c11"></anchor>

Dechaume, M., Mercier, M., Feybesse, C., Lubart, T., Chouvelon, G., Kermarrec, S., & Tordjman, S. (2024). Is intelligence necessary and sufficient for creativity? An analysis of convergent and divergent thinking. Learning and Individual Differences, 116, Article 102575. 10.1016/j.lindif.2024.102575

<anchor name="c12"></anchor>

Duncker, K. (1945). On problem-solving (Psychological Monographs, No. 270.).

<anchor name="c13"></anchor>

Ericsson, K. A., & Simon, H. A. (1980). Verbal reports as data. Psychological Review, 87(3), 215–251. 10.1037/0033-295X.87.3.215

<anchor name="c14"></anchor>

Eymann, V., Beck, A. K., Jaarsveld, S., Lachmann, T., & Czernochowski, D. (2022). Alpha oscillatory evidence for shared underlying mechanisms of creativity and fluid intelligence above and beyond working memory-related activity. Intelligence, 91, Article 101630. 10.1016/j.intell.2022.101630

<anchor name="c15"></anchor>

Eymann, V., Beck, A. K., Lachmann, T., Jaarsveld, S., & Czernochowski, D. (2024). Reconsidering divergent and convergent thinking in creativity—A neurophysiological index for the convergence-divergence continuum. Creativity Research Journal, 1–8. 10.1080/10400419.2024.2419751

<anchor name="c16"></anchor>

Eysenck, H. J. (2003). Creativity, personality and the convergent-divergent continuum. In M. A.Runco (Ed.), Critical creative processes (pp. 95–114). Hampton Press. <a href="https://api.semanticscholar.org/CorpusID:221272407" target="_blank">https://api.semanticscholar.org/CorpusID:221272407</a>

<anchor name="c17"></anchor>

Finke, R. A., Ward, T. B., & Smith, S. M. (1992). Creative cognition: Theory, research, and applications. MIT Press. 10.7551/mitpress/7722.001.0001

<anchor name="c18"></anchor>

Hommel, B., Colzato, L. S., Fischer, R., & Christoffels, I. K. (2011). Bilingualism and creativity: Benefits in convergent thinking come with losses in divergent thinking. Frontiers in Psychology, 2, Article 273. 10.3389/fpsyg.2011.00273

<anchor name="c19"></anchor>

Jacobs, W. J., & Metcalfe, J. (2024). The role of Curiosity 1 and Curiosity 2 in the emergence of insight. In C.Salvi, J.Wiley, & S. M.Smith (Eds.), Emergence of insight (pp. 155–139). Cambridge University Press. 10.1017/9781009244244.006

<anchor name="c20"></anchor>

Jacoby, L. L. (1991). A process dissociation framework: Separating automatic from intentional uses of memory. Journal of Memory and Language, 30(5), 513–541. 10.1016/0749-596X(91)90025-F

<anchor name="c21"></anchor>

Jung-Beeman, M., Bowden, E. M., Haberman, J., Frymiare, J. L., Arambel-Liu, S., Greenblatt, R., Reber, P. J., & Kounios, J. (2004). Neural activity when people solve verbal problems with insight. PLOS Biology, 2(4), Article e97. 10.1371/journal.pbio.0020097

<anchor name="c22"></anchor>

Kim, K. H., & Pierce, R. A. (2013). Convergent versus divergent thinking. In E. G.Carayanis (Ed.), Encyclopedia of creativity, invention, innovation and entrepreneurship (pp. 245–250). Springer. 10.1007/978-1-4614-3858-8_22

<anchor name="c23"></anchor>

Kizilirmak, J. M., Galvao Gomes da Silva, J., Imamoglu, F., & Richardson-Klavehn, A. (2016). Generation and the subjective feeling of “aha!” are independently related to learning from insight. Psychological Research, 80(6), 1059–1074. 10.1007/s00426-015-0697-2

<anchor name="c24"></anchor>

Knoblich, G., Ohlsson, S., & Raney, G. E. (2001). An eye movement study of insight problem solving. Memory & Cognition, 29(7), 1000–1009. 10.3758/BF03195762

<anchor name="c25"></anchor>

Knoblich, G., Öllinger, M., & Spivey, M. J. (2005). Tracking the eyes to obtain insight into insight problem solving. Cognitive Processes in Eye Guidance, 355–375. 10.1093/acprof:oso/9780198566816.003.0015

<anchor name="c26"></anchor>

Kohn, N. W., & Smith, S. M. (2011). Collaborative fixation: Effects of others’ ideas on brainstorming. Applied Cognitive Psychology, 25(3), 359–371. 10.1002/acp.1699

<anchor name="c27"></anchor>

Kounios, J., & Beeman, M. (2009). The Aha! moment: The cognitive neuroscience of insight. Current Directions in Psychological Science, 18(4), 210–216. 10.1111/j.1467-8721.2009.01638.x

<anchor name="c28"></anchor>

Kvavilashvili, L., & Mandler, G. (2004). Out of one’s mind: A study of involuntary semantic memories. Cognitive Psychology, 48(1), 47–94. 10.1016/S0010-0285(03)00115-4

<anchor name="c29"></anchor>

Madore, K. P., Addis, D. R., & Schacter, D. L. (2015). Creativity and memory: Effects of an episodic-specificity induction on divergent thinking. Psychological Science, 26(9), 1461–1468. 10.1177/0956797615591863

<anchor name="c30"></anchor>

Maier, N. R. (1931). Reasoning in humans. II. The solution of a problem and its appearance in consciousness. Journal of Comparative Psychology, 12(2), 181–194. 10.1037/h0071361

<anchor name="c31"></anchor>

Metcalfe, J. (1986). Premonitions of insight predict impending error. Journal of Experimental Psychology: Learning, Memory, and Cognition, 12(4), 623–634. 10.1037/0278-7393.12.4.623

<anchor name="c32"></anchor>

Metcalfe, J., & Wiebe, D. (1987). Intuition in insight and noninsight problem solving. Memory & Cognition, 15(3), 238–246. 10.3758/BF03197722

<anchor name="c33"></anchor>

Nakagawa, S., Johnson, P. C. D., & Schielzeth, H. (2017). The coefficient of determination R2 and intra-class correlation coefficient from generalized linear mixed-effects models revisited and expanded. Journal of the Royal Society, Interface, 14(134), Article 20170213. 10.1098/rsif.2017.0213

<anchor name="c34"></anchor>

Nijstad, B. A., De Dreu, C. K., Rietzschel, E. F., & Baas, M. (2010). The dual pathway to creativity model: Creative ideation as a function of flexibility and persistence. European Review of Social Psychology, 21(1), 34–77. 10.1080/10463281003765323

<anchor name="c35"></anchor>

Patterson, K. E., Meltzer, R. H., & Mandler, G. (1971). Inter-response times in categorized free recall. Journal of Verbal Learning and Verbal Behavior, 10(4), 417–426. 10.1016/S0022-5371(71)80041-5

<anchor name="c36"></anchor>

Roediger, H. L., & Thorpe, L. A. (1978). The role of recall time in producing hypermnesia. Memory & Cognition, 6(3), 296–305. 10.3758/BF03197459

<anchor name="c37"></anchor>

Rohrer, D., & Wixted, J. T. (1994). An analysis of latency and interresponse time in free recall. Memory & Cognition, 22(5), 511–524. 10.3758/BF03198390

<anchor name="c38"></anchor>

Runco, M. A., & Acar, S. (2012). Divergent thinking as an indicator of creative potential. Creativity Research Journal, 24(1), 66–75. 10.1080/10400419.2012.652929

<anchor name="c39"></anchor>

Salvi, C., & Bowden, E. (2024). Insight, what happens backstage? In C.Salvi, J.Wiley, & S. M.Smith (Eds.), Emergence of insight (pp. 280–308). Cambridge University Press. 10.1017/9781009244244.013

<anchor name="c40"></anchor>

Salvi, C., Bricolo, E., Franconeri, S. L., Kounios, J., & Beeman, M. (2015). Sudden insight is associated with shutting out visual inputs. Psychonomic Bulletin & Review, 22(6), 1814–1819. 10.3758/s13423-015-0845-0

<anchor name="c41"></anchor>

Salvi, C., Bricolo, E., Kounios, J., Bowden, E., & Beeman, M. (2016). Insight solutions are correct more often than analytic solutions. Thinking & Reasoning, 22(4), 443–460. 10.1080/13546783.2016.1141798

<anchor name="c42"></anchor>

Schooler, J. W., & Melcher, J. (1995). The ineffability of insight. In S. M.Smith, T. B.Ward, & R. A.Finke (Eds.), The creative cognition approach (pp. 97–133). The MIT Press.

<anchor name="c44"></anchor>

Schooler, J. W., Ohlsson, S., & Brooks, K. (1993). Thoughts beyond words: When language overshadows insight. Journal of Experimental Psychology: General, 122(2), 166–183. 10.1037/0096-3445.122.2.166

<anchor name="c43"></anchor>

Shen, W., Hommel, B., Yuan, Y., Chang, L., & Zhang, W. (2018). Risk-taking and creativity: Convergent, but not divergent thinking is better in low-risk takers. Creativity Research Journal, 30(2), 224–231. 10.1080/10400419.2018.1446852

<anchor name="c45"></anchor>

Shiffrin, R. M. (1970). Memory search. Models of Human Memory, 375–447.

<anchor name="c46"></anchor>

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. 10.1037/1931-3896.2.2.68

<anchor name="c48"></anchor>

Smith, S. M., & Beda, Z. (2024). The past and future of research on so-called ‘incubation’ effects. In C.Salvi, J.Wiley, & S. M.Smith (Eds.), Emergence of insight (pp. 13–38). Cambridge University Press. 10.1017/9781009244244.002

<anchor name="c49"></anchor>

Smith, S. M., Gerkens, D. R., & Angello, G. (2017). Alternating incubation effects in the generation of category exemplars. The Journal of Creative Behavior, 51(2), 95–106. 10.1002/jocb.88

<anchor name="c50"></anchor>

Tellez, M. O., Kenett, Y., Benedek, M., Hills, T., Beranger, B., Lopez-Persem, A., Bieth, T., & Volle, E. (2024). Switching, fast and slow: Deciphering the dynamics of memory search, its brain connectivity patterns, and its role in creativity. Preprint, pp. 1–34. 10.21203/rs.3.rs-3826172/v1

<anchor name="c51"></anchor>

Zabelina, D. L., & Robinson, M. D. (2010). Creativity as flexible cognitive control. Psychology of Aesthetics, Creativity, and the Arts, 4(3), 136–143. 10.1037/a0017379

Submitted: January 17, 2025 Revised: September 28, 2025 Accepted: October 11, 2025

*