Eugen Fischer, Keith Allen and Paul Engelhardt
Affiliation: University of East Anglia, University of York, University of East Anglia
Category: Philosophy
Keywords: folk conceptions of vision, Direct and Indirect Realism, belief fragmentation, naive vs scientific theories, appeals to common sense
Date: Tuesday 2nd of September
Time: 14:30
Location: Gen. Henryk Dąbrowski Hall (006)
View the full session: Metaphysics of Perceptual Experience
Philosophical debates about ‘aporetic’ problems offer a window on conflicts within ‘common sense’. While persistent intra-personal conflicts between co-represented naïve and scientific theories are well documented (e.g., Shtulman & Legare, 2020; Shtulman & Valcarcel, 2012; cf. Barlev et al., 2018), clashes between pre-scientific folk beliefs have received little empirical attention, despite their relevance to both psychology and philosophy: In psychology, they have major implications for accounts of human rationality and cognitive architecture, providing key evidence that beliefs are stored in causally independent data structures (Bendaña & Mandelbaum, 2021; Leiser, 2001; Sommer et al., 2023). In philosophy, conflicts between pre-scientific folk beliefs suggest that ‘common sense’ is too conflicted to merit the epistemic default status it is accorded in many debates.
We address this important gap by examining the prevalence, and the naïve or scientific status of conflicting beliefs about vision. Direct Realism (DR) maintains that physical objects are ‘directly’ present in perception (Lyons, 2023). Indirect Realism (IR) maintains we see physical objects only by being aware of mental images or ‘sense-data’ (e.g., Jackson, 1977; Robinson, 1994; Russell, 1912/2000). Philosophical debates about the aporetic ‘problem of perception’ (review: Crane & French, 2021) typically take for granted that there is such a thing as ‘the’ common-sense conception of vision, which is consistent with DR. This assumption shapes the methodology of those debates: Many contributors maintain that, in virtue of capturing common sense, DR should be accepted in the absence of good reasons to the contrary (review: Genone, 2016), while IR requires an error theory that explains how common sense could go wrong (e.g., Russell, 1912/2000; Boghossian & Velleman, 1989).
We review three agreement-rating studies which have validated a new belief inventory that assesses core metaphysical and epistemological beliefs associated with DR and IR, respectively (Anonymized). Cluster analyses reveal that both conceptions are endorsed by 20-40% across lay samples (N=100, N=500, N=320). Laypeople are clearly familiar with both views, and many endorse both. Exploratory factor analyses provide evidence of at least two distinct data structures that drive responses to DR and IR items, respectively, suggesting conflicting conceptions of vision are co-represented in lay minds.
Our talk asks whether these conflicting conceptions are treated as naïve or scientific, in lay cognition, where naïve theories get suppressed in favour of conflicting theories deemed more scientific (Shtulman & Valcarcel, 2012).
Philosophical orthodoxy views DR as naïve and IR as more scientific. Together with findings from the new belief inventory, this entails: H Laypeople regard IR as more scientifically accurate than DR.
We present an alternative account that regards both conceptions as pre-scientific: DR is ultimately grounded in experiential world knowledge about vision events (Rumelhart, 1978), and IR in a Cartesian Theatre conception (Forstmann & Burgmer, 2022) that is due to partial access to an implicit model of endogenous attention (Graziano, 2022). Our account suggests: H* Laypeople regard neither IR nor DR as scientifically accurate.
A scientific accuracy rating task assessed H against H*.
Methods 200 participants (UK residents, aged 18-80, 43% male), recruited via Prolific, had no higher than undergraduate degree, and none in sciences, psychology, or philosophy. They used a 7-point scale from “I am highly confident that this statement is scientifically inaccurate” to “I am highly confident that this statement is scientifically accurate”, to rate 24 statements. 8 statements articulated key claims of DR and IR, using a concrete example, e.g.: PD When you look at a bus, you see just the bus and not a mental image of the bus. PI When you look at a tomato, you see a mental image of the tomato and not just the tomato. PCD When you look at a fork, you see just the fork and not a mental image caused by the fork. 8 articulated the same claims in more abstract terms that may sound more scientific, e.g.: PD-A When you look at a physical object, you see just the object and not a mental image of the object. 8 fillers afforded a manipulation check. We manipulated Realism (Direct vs Indirect) and Abstraction (Concrete vs Abstract) in a 2x2 within-subject design.
Results Repeated-measures ANOVAs revealed a significant interaction (F(1,195)=37.9, p.001, η2=.16) (Figure 1) and a small effect of Realism (F(1,195)= 4.93, p=.028, η2 =.03), but no abstract-concrete effect (p=.85). In the abstract condition, mean ratings for DR and IR were virtually identical and at mid-point; in the concrete condition, mean ratings for DR were just above mid-point and just below for IR (Figure 1). The means hid considerable individual variation. However, no metaphysical statement of either conception was judged scientifically accurate by even half the sample (Table 1), and all DR-IR frequency comparisons remained insignificant (all Chi21.98, well shy of the critical Chi2 value (df=1) of 3.84). These findings favour our hypothesis H* that neither DR nor IR are generally regarded as scientifically accurate.
Next, we assessed to what extent DR could, even so, function, in lay cognition, as a naïve theory that gets suppressed in favour of an IR conception deemed more scientific. As a minimum, this requires individuals to deem IR more scientifically accurate than DR. More plausibly, it requires that individuals deem IR scientific, but not DR. The relevant proportions remained low (Table 2). The proportion of participants who meet the more plausible criterion (31.6%) is almost identical with the proportion of participants whose mean IR ratings are clearly (1 point) above their mean DR ratings (30.6%). We infer against philosophical orthodoxy that, for at most a third of participants, DR potentially functions as a naïve theory suppressed in favour of a more scientific IR theory.
Fillers provided a successful manipulation check: Statements consistent and inconsistent with familiar scientific theories attracted means significantly above and below midpoint, respectively (p’s.001).
Findings suggest that there is no such thing as ‘the’ common-sense conception of vision: At least two conflicting conceptions, consistent with DR and IR, respectively, are co-represented in laypeople’s minds; and these do not generally function as naïve vs scientific theory but have equal claim to being ‘common sense’.