Follow-up_Moral Language and Moore's Paradox. Challenging Expressivism
Year: 2025
Type: study-registration
Abstract: In a recent empirical study, Sorio-Ruiz and Franzen (2022, https://osf.io/568cz) investigate what Schroeder has called the Parity Thesis: '[M]oral sentences are related to noncognitive, desire-like states of mind in the same way that ordinary descriptive sentences are related to ordinary beliefs—they express them.' (Schroeder, 2008, p. 3). For instance, it is widely accepted that utterances of the form "It is raining outside" express the speaker's belief that it is raining outside. As a consequence, it is also widely accepted that denying this inference yields an infelicitous statement, called a Moorean Incoherence: # (1) "It is raining outside, but I don't think it is raining outside." According to expressivism about moral language, moral terms like "wrong" express the speaker's non-cognitive attitude, e.g. their disapproval. Here is an example, taken from Sorio-Ruiz and Franzen (2022): "Murder is wrong" expresses the speaker's disapproval. According to the Parity Thesis, the following sentence should sound just as infelicitous as (1) ? (2) "Murder is wrong, but I don't disapprove of it." In their empirical research, Sorio-Ruiz and Franzen investigate whether (2) sounds, indeed, infelicitous to ordinary speakers and whether they are infelicitous to the same degree as (1). They formulate the following two Hypotheses: Hypothesis (I): First-personal sentences are significantly less acceptable than their third-personal counterparts. That is, we predict a main effect of Person; Hypothesis (II*): The difference in acceptability between first- and third-personal sentences is smaller for 'disapproval'-sentences than for 'think'-sentences. That is, we predict an interaction between Person and Attitude. Contrast (1) with (3) # (1) "It is raining outside, but I don't think it is raining outside." ! (3) "It is raining outside, but Ann doesn't think it is raining outside." In line with H1, while (1) is infelicitous, (3) is not. Now, consider (2) and (4) ? (2) "Murder is wrong, but I don't disapprove of it." ? (4) "Murder is wrong, but I don't think it is wrong." Intuitively, (4) sounds more acceptable than (2). Again, if the data supported this intuition, this would count in favor of H1. According to H2, though, the difference between (1) and (3) would be larger than between (2) and (4). In the rest of this preregistration, we are only concerned with Experiment 1. Sorio-Ruiz and Franzen do find partial support for the Parity thesis. In this study, we wish to address three closely related limitations of Sorio-Ruiz and Franzen's research: First, the authors use only one moral term, namely "wrong." While their results seem clear for this term, it is uncertain whether the effect extends to other moral terms as well. Second, "wrong" is a prime example of what many philosophers call thin concepts or "mere evaluative." It is a purely evaluative concept with no (or extremely little) descriptive content. However, one may wonder whether thick concepts like "rude," "cruel," or "manipulative" work alike. Third, the authors limited themselves to a negative term, leaving aside the positive analogue "right." Recent empirical evidence by Willemsen & Reuter (2021), Baumgartner, Willemsen, & Reuter (2022), and Willemsen, Baumgartner, Cepollaro, & Reuter (2024) yet demonstrates that positive and negative moral terms behave quite differently in experimental settings, suggesting that the evaluative content is conveyed in different ways. To overcome these limitations, we extended Sorio-Ruiz and Franzen's original study and further added: - a list of three negative thick concepts - a list of closely matched positive counterparts to these negative thick concepts - the positive counterpart to "wrong," namely "right". This design was tested in a preregistered study related to this project. In this follow-up, we test two things: First, is there an effect of gender, such that the speaker's name affects how felicitous their statement is considered to be? In our first experiment, the speaker's name was Sally. While we do not expect any changes if the speaker's name were substituted with the name Tom, we still wish to test this possibility. We only test first-person statements, where the speaker voices their own opinion. Second, why do people in our original study judge moralized versions of Moore's paradox to be so inconsistent? While almost 95% of participants judged traditional Moorean paradoxes to "not make sense" in first-person formulations ("It is raining but I don't think it is raining"), only 59 to 75% (depending on condition) did so when the formulation was moralized (e.g., "Cyberbulling is wrong but I don't think it is wrong").
Source: Open Science Framework
Authors: Willemsen, Pascale, Baumgartner, Lucien
Language: English
Subfield: Philosophy
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7147424354