Andreas Arslan and Jonathan Kominsky
Affiliation: Central European University PU
Category: Psychology
Keywords: imagination, mental imagery, event simulation
Date: Thursday 4th of September
Time: 18:30
Location: Maria Skłodowska-Curie Hall (123)
View the full session: Imagination
The language used to describe the workings and products of imagination often likens it to vision: We talk of mental images, picturing situations in our head, and seeing people or objects in our mind's eye. Are these merely convenient metaphors? Or is to imagine, at least occasionally, to mentally simulate scenes in a format, and with a vividness, that parallels direct experience?
Naturally, philosophers have long wondered about the nature of mental imagery. Descartes used the example of a 'Chiliagon,' a thousand-sided polygon, to illustrate the difference between mentally conceiving of an object and literally envisioning it (Descartes, 2013). Much later, Ryle roundly rejected the notion of mental images that hang suspended in some internal 'gallery' (Ryle, 2009). By contrast, Block argued that the occasional imperfection or crudeness of mental images is not evidence of them being non-pictorial, calling the claim that it is the 'photographic fallacy' (Block, 1983). In neuroscience and cognitive science, recent theoretical proposals that the episodic memory system can be viewed as a general-purpose event simulation system (Addis, 2020) and that episodic memories are those event simulations that have been metacognitively tagged as memories (Mahr et al., 2023) suggest imagined experiences have the potential to be quasi-perceptual in character. This conflicts with the findings of Bigelow et al. (2023), who instructed participants to imagine fairly short sentences as vividly as they could: In their experiments, features of scenes that would be difficult to overlook in real-life situations often were not imagined by participants.
Building on the methods employed by Bigelow et. al (2023), we investigated how complete (similar to direct perception) and coherent (free of contradictions) people's imagined scenes are. The four online experiments we conducted (Experiment 1; Experiment 2a – 2c) all shared a similar a design centered around presenting short narratives to participants and asking them to imagine them as vividly as they could.
In Experiment 1, we investigated the relationship between the completeness and coherence of imagined scenes. To this end, we asked participants to read and vividly imagine a 72-words-long riddle-like text (adapted from Bar-Hillel et al., 2018) for 60 seconds. We chose a riddle because they are often ambiguous and apparently contradictory – ‘solving the riddle,' in these cases, is equivalent to finding a way to imagine the described scene without contradictions. This gave us a binary measure of whether participants had imagined the scene coherently (7 out of 38, or 18.4%, did). To assess completeness, we sequentially asked participants whether they had imagined certain basic features (nine in total) of the scene. A feature was presented to them textually (for example, Landscape or Road width) and they had to state whether it was 'part of their mental image' or not. Paralleling Bigelow et al.'s (2023) results, participants' mental scenes frequently were far from complete, the average number of imagined features being M = 4.87 (SD = 1.70) out of 9 possible ones. In agreement with our preregistered hypothesis, a point-biserial correlation analysis revealed that there was no significant relation, rpb(36) = 0.02, p = 0.91, between the number of features imagined and whether a participant had solved the riddle, indicating that coherence does not imply completeness and vice versa.
Experiments 2a – 2c inverted the approach of Experiment 1. Instead of presenting an ambiguous text to participants and testing their ability to imagine it coherently, we used two vignettes (191 and 211 words long) that each contained an objective spatial contradiction. In Vignette 1, this took the form of contradictory descriptions of the location of furniture in a room and in Vignette 2 it concerned the position of the sun in the sky and consequent direction of a shadow.
Experiment 2a followed a simple methodology: Participants had to imagine one of the vignettes and were then asked whether they noticed an 'objective error' in the story. If they answered Yes, they had to specify what the error was. Many participants did not spot the error (7 out of 15 were successful in Vignette 1; only 2 out of 15 in Vignette 2).
In Experiment 2b we tested whether people would have more success in finding the error in Vignette 2 if visual material scaffolded their imagination. While participants in one condition had to point out the error in writing as in Experiment 2a, those in another condition were shown a map of the locations that had been described in the text. The map did not improve participants' performance: While 2 out of 23 participants in the text condition found the error, none (0 out of 17) were successful in the map condition.
In Experiment 2c we investigated whether an individual's mental rotation ability predicts if they manage to detect a spatial contradiction in a vignette. Would those who responded more quickly or made fewer errors in a mental rotation task (we used stimuli from Ganis & Kievit, 2015) also be more likely to correctly point out the error? In this final experiment, we only presented Vignette 1 to participants, as just a small proportion of participants had given correct responses to Vignette 2 in previous experiments. We found no significant association between accuracy on the mental rotation task and error detection, rpb(28)= 0.07, nor between average response time and error detection, rpb(28)= 0.03, p = 0.86.
Taken together, the data we have gathered so far suggest that the scenes people imagine when reading a text are not just a few missing details away from being simulacra of reality. Instead of an integrated scene where all relations between parts are specified or at least inferable, mental images may frequently be composed of multiple, somewhat independent, feature patches that together create the illusory impression of an organized whole.
OSF repository Exp1: https://osf.io/prkn5/?view_only=156a1413f1b844f69c0ff3a58a737ff4
OSF repository Exp 2a – 2c: https://osf.io/379zx/files/osfstorage?view_only=4787509bbd4c4d2aab8700ee691dbfc7
Preregistration Exp. 1: https://osf.io/r6fyg/?view_only=ca05a8b9f01440f6806aaec799ea6355
Preregistration Exp. 2c: https://osf.io/yq6f5/?view_only=54dcbb102f9a4d69920a21cb24071780
Addis, D. R. (2020). Mental time travel? A neurocognitive model of event simulation. Review of Philosophy and Psychology, 11(2), 233-259. https://doi.org/10.1007/s13164-020-00470-0.
Bar-Hillel, M., Noah, T., & Frederick, S. (2018). Learning psychology from riddles: The case of stumpers. Judgment and Decision Making, 13(1), 112-122. https://doi.org/10.1017/S193029750000886X
Bigelow, E. J., McCoy, J. P., & Ullman, T. D. (2023). Non-commitment in mental imagery. Cognition, 238, 105498. https://doi.org/10.1016/j.cognition.2023.105498
Block, N. (1983). The photographic fallacy in the debate about mental imagery. Nous, 651-661.
Descartes, R. (2013). Meditations on first philosophy. Broadview Press.
Ganis, G., & Kievit, R. A. (2015). A New Set of Three-Dimensional Shapes for Investigating Mental Rotation Processes: Validation Data and Stimulus Set. Journal of Open Psychology Data, 3(1), e3. https://doi.org/10.5334/jopd.ai
Mahr, J. B., van Bergen, P., Sutton, J., Schacter, D. L., & Heyes, C. (2023). Mnemicity: A Cognitive Gadget? Perspectives on Psychological Science, 18(5), 1160-1177. https://doi.org/10.1177/17456916221141352
Ryle, G., & Tanney, J. (2009). The concept of mind. Routledge.