Researchers involved: Lennart Fritzsche (Goethe University Frankfurt) & Sebastian Walter (Visual and Non-visual Means of Perspective Taking in Language)
Negation is a key component of human language, intrinsically tied to cognitive processes that allow us to reason about what is not the case—a distinctive property of human communication. As such, negation not only encodes a well-defined grammatical function, but also intersects with a variety of grammatical and cognitive domains. While there has been extensive research on negation in spoken and written language, non-verbal means used to mark negation have been investigated scarcely (but see, e.g., González-Fuente et al., 2015; Li et al., 2016). The experiments outlined here aim to fill this gap by examining how gestural markers of negation compare with standard verbal response particles.
This collaboration builds directly on the successful outcomes of a previous short-term collaboration (described below) conducted by the same authors. ViCom stands to gain substantially from this joint project: In addition to fostering collaboration between two theoretically oriented early-career researchers—one from the ViCom project Visual and non-visual means of perspective taking in language, Frankfurt/Hamburg (Sebastian Walter) and one ViCom-affiliated MA student from Goethe University Frankfurt (Lennart Fritzsche)—the findings promise to lay critical groundwork for broader multimodal theory development, one of ViCom’s main objectives.
German employs a threefold response particle system. In principle, when speakers affirm negative antecedents, they can choose either ja (‘yes’) or nein (‘no’) (Claus et al., 2017). Beyond spoken forms, response elements also appear in the visual domain—here, we focus on headnods and headshakes used pro-speech, i.e., in their speech-replacing use. Loos and Repp (2024) observe that German speakers produce more pro-speech headnods than headshakes when affirming or rejecting negative antecedents. However, these pro-speech head gestures were comparatively sparse in the data, preventing a comprehensive account of their use and interpretation.
Despite this limited evidence, it appears plausible to assume that headnods and headshakes may pattern with ja and nein, respectively. Nonetheless, when gestures co-occurred with spoken responses, Loos and Repp (2024) report more headnods than headshakes co-occurring with doch—the third German response particle, which has to be used when rejecting negative antecedents. When responding to negative polar questions, we therefore hypothesized that headshakes behave like ja and nein. Headnods, on the other hand, might display an additional doch-like interpretation, suggesting a broader range of usage.
To test this, we conducted a forced-choice experiment in German as part of our first short-term collaboration. Participants (n = 592) first read a short context that licensed a negative polar question, as in (1-a). Then, they saw a videotaped discourse in which a male speaker uttered (1-a), followed by one of four responses, cf. (1-b).
(1)
A: Sind die Schuhe nicht im Kleiderschrank? (‘Are the shoes not in the closet?’)
B1: Ja. (‘Yes.’)
B2: Nein. (‘No.’)
B3: HEADNOD
B4: HEADSHAKE
Afterwards, participants were shown two images: one depicting shoes in the closet and one depicting no shoes in the closet. They had to select the image that best matched B’s response. The experiment thus employed a single-factor design with four levels (RESPONSE ELEMENT: ja vs. nein vs. headnod vs. headshake). A logistic regression model revealed a main effect of RESPONSE ELEMENT. Pairwise comparisons showed no significant difference between ja and nein, nor between nein and headshake. The headnod condition, however, diverged significantly from ja and thus from all other conditions. These findings suggest that, in negative contexts, headnods can serve two functions: (i) affirming negative antecedents (preferred) and (ii) mirroring a doch-like rejection of negative antecedents.
In our follow-up studies planned within this short-term collaboration, we seek to establish whether the doch-like interpretation of headnods can be boosted by manipulating eyebrow movement and nod amplitude. First, we will replicate the experiment with ja, headnod, and doch to see how headnod falls between the choice proportions of ja (and thus nein/headshake) and doch. This experiment serves as a baseline for manipulating headnods in the next experiment, as we need a reference point to determine exactly where the choice proportions for headnod lie between ja and doch. A main effect for RESPONSE ELEMENT is predicted. 450 participants will be recruited.
A second experiment will test neutral headnod (headnod in the previous studies), high-amplitude headnod, headnod with brow furrowing, and doch. Again, a main effect for RESPONSE ELEMENT is predicted. Specifically, while the doch condition should be associated with the highest choice proportions for rejection, and neutral headnod with the highest proportions for affirmation, the modified conditions high-amplitude headnod and headnod with brow furrowing should lie in-between the conditions, i.e., leaning more towards doch as we hypothesize brow furrowing and intensified headnods to be markers of rejection. This hypothesis is derived from the observation that there is generally more gestural marking in rejections and specifically, there is more movement of the eyebrows in rejections of negative antecedents (Loos and Repp, 2024). 600 participants will be recruited.