Language

Causal explanations and causal concepts are often regarded as being domain-specific across languages and cultures. Accordingly, a large amount of psychological studies scrutinized causal inferences within the boundaries of predefined domains, such as biology, psychology or physics, and some studies even hint at the availability of meta-representations that include abstract knowledge about the structures of causal relationships in a specific domain. To what extent, however, are causal judgments influenced by domain-specific information in different languages and to what degree do assumptions on causal structures differ between domains? Several experimental studies will be presented that have investigated the impact of domain-specific information on causal judgments using different task in multiple cultural or linguistic contexts. Results indicate that causal knowledge and causal judgments vary between domains and that these variations seem to be roughly similar across the languages investigated. Thus, the results provide interesting implications for discussions on the cross-linguistic unity or diversity of causal cognition.

Moral judgment research frequently uses sacrificial dilemmas, in which one has to decide whether to kill one person in order to save several others. In this paradigm, not killing is usually assumed to suggest adherence to absolute norms, while killing is taken to indicate sensitivity to consequences of an action. Research employing this approach indicates the existence of a foreign language effect (FLE), such that willingness to engage in sacrificial killing is higher when scenarios are considered in a foreign compared to one’s native language. However, methodological limitations prevent much of past research from speaking clearly about its underlying mechanisms. Moreover, as some evidence suggests the FLE may be restricted to scenarios enforcing high levels of personal involvement, the boundary conditions of this effect are somewhat unclear. Avoiding these limitations we applied multinomial processing tree modeling to estimate sensitivity to aggregate consequences, norm-sensitivity, and preference for inertia over interference as independent determinants of dilemma responses. Results of two experiments (N1 = 247, N2 = 574) suggest that the FLE may be restricted to high-involvement dilemmas, in which foreign language reduces norm-sensitivity and inertia alike. In response to low-involvement dilemmas no consistent effects were observed. These findings thus help to clarify how several cognitive mechanisms jointly contribute to the FLE in dilemma-judgment. While partly compatible with previous research, our findings also suggest the FLE to be in part an artefact resulting from uncontrolled response tendencies, specifically inertia, which systematically bias responses in other dilemma paradigms.

Switching between two different tasks is facilitated by knowing which task needs to be performed next. However, we do not yet know how predictability of sequences affects language switching. Given that the costs of task switching and language switching are correlated in individuals (Declerck, Grainger, Koch & Philipp, 2017, JML), we hypothesized that language switching—like task switching—would benefit from predictability. To investigate this, unbalanced German-English bilinguals (L1 = German; L2 = English) were recruited in two independent experiments (N1 = 22, N2 = 35). In both, participants named images of frequent semantic concepts either in German or in English. In the majority of blocks, trials followed a predictable language sequence (L1, L1, L2, L2, L1, L1, etc.), while the language sequence was random in a negative-transfer block. In both experiments, we observed shorter reaction times in German than English naming trials. In addition, switch trials resulted in longer RTs than repetition ones. Importantly, however, we found no evidence for a predictability benefit in language switching; in contrast, there was a tendency for a predictability cost in Experiment 1. Together, our results suggest that predictable language sequences do not facilitate switching between languages, but that instead knowing which language comes up next increased cross-lingual competition, potentially due to larger parallel activation of L1 and L2.

In the present study, we examined the influence of morphological configuration (i.e., the structure of morphemes in a word) in language switching. We conducted several experiments in which not only the relevant language (German, English or Spanish) varied between trials but also the rule for the morphological configuration of a word. In Experiment 1 and 2, we used two-digit numbers because the morphological configuration (i.e., the composition rule of the number words) varies between languages (e.g. inverted composition rule with unit before decade vs. non-inverted composition rule with decade before unit). Experiment 1, participants (n=36) had to name a visually presented two-digit number, whereas the numbers were presented auditorily in Experiment 2 and participants (n=48) had to determine the numerical distance (i.e. 2, 3 or 4) between the numbers in the current and the previous trial. In Experiment 3, participants (n=48) had to name compound words for which the morphological configuration (verb+noun vs. noun+verb) differed between languages. The results of all experiments demonstrated language-switch costs, which is a better performance in language-repetition trials than in language-switch trials. Importantly, language-switch costs were modulated by repeating or switching the morphological configuration. More specifically, a language-repetition benefit (i.e., switch cost) was observed when the morphological configuration repeated in two successive trials but was reduced or even reversed when the morphological configuration switched. Our results indicate that the morphological configuration plays a critical role in the representation and the production of compound words in language switching.