Electrophysiological evidence for a Whorfian double dissociation of categorical perception across two languages

Abstract

Languages vary considerably in how they group objects into categories. For example, the word taza in Spanish can refer to either cup or mug in English, whereas glass can refer to either copa or vaso – two different types of glasses –, in Spanish. It is still debated whether such language distinctions cause differences in early perceptual processing between speakers of different languages. Here, for the first time, we test the effects of terminology on pre-attentive indices of categorical perception symmetrically across two languages. In a visual mismatch negativity (vMMN) paradigm, native speakers of English or Spanish were flashed with different exemplars from four object categories: cup, mug, copa, and vaso. We measured the vMMN elicited by language-dependent within- and cross-category changes in the peripheral visual field. Results showed significant vMMN modulations for categorical contrasts distinguished by participants’ native language, but not for objects grouped under the same label. P3a modulations showed the reverse pattern, with significant modulations only for those categorical contrast unmarked in the native language. These results provide the strongest evidence to date that native language labels influence categorical perception of objects at pre-attentive stages of perceptual encoding. In addition, they show that non-verbally encoded categorical dissociations also reorient attention but at a later stage and in a more controlled fashion. Therefore, we conclude that perceptual encoding is shaped by one’s language.

An Introduction to the Cognitive Neuroscience of Language Embodiment and Relativity Special Issue of the Language Learning Cognitive Neuroscience Series

“…The authors found a right visual field categorical perception effect both immediately after learning and after consolidation, affecting the amplitude of early peaks of ERPs (P1, N1, and N2), elicited by a circular display of objects featuring either a within-or cross-category target. This result provides strong support for a causal role of language terminology in early categorical perception, since it was obtained in a learning paradigm when participants had no preexisting knowledge of the labels or the function of the objects presented (a fundamental difference from participants tested by Casaponsa et al, 2024).…”

“…In an oddball design manipulating the degree of matching between prime animation and target picture, the authors found that target pictures matching the animation in terms of object shape, but not transitivity, elicited a reduced P300 effect as compared to pictures matching the animation in terms of transitivity but not object shape (the P300 indexes spontaneous shifts of attention towards an infrequent, salient stimulus). A further test involving an oddball paradigm (following the same principle as that used by Casaponsa et al, 2024) showed that transitivity deviants elicited a vMMN modulation whereas object shape deviants did not. The authors concluded that implicit learning of a novel grammatical morpheme can immediately alter preattentive visual processing.…”

“…The next four papers move from behavioral evidence to implicit measures of brain activity, namely event-related potentials (ERPs), less likely to be modulated in a strategic manner: ERPs provide indices of visual perception and unconscious information processing. Casaponsa et al (2024) presented native speakers of Spanish and native speakers of English with pictures of cups, mugs, glasses with a stem, and glasses without stems (regular glasses) in an oddball paradigm featuring one object type frequently (standard) and another rarely (deviant). The main (irrelevant) task was to detect a change in the appearance of the fixation cross presented in the center of the screen, while differences between objects only concerned peripheral (unattended) vision.…”

“…Critically, in all the studies reported in this special issue, the authors distanced themselves from the language representations or language mechanisms hypothetically driving the phenomenon being measured, by using tasks in which language has minimal involvement or relevance: smell/taste categorization (Bylund et al, 2024), color discrimination (Sinkeviciute et al, 2024), time conceptualization (Athanasopoulos & Su, 2024), and temporal transition estimation (Vanek & Zhang, 2024). Another set of studies not only used tasks keeping language involvement peripheral, but also collected implicit measures of unconscious processing directly derived from brain activity that are not susceptible to subjective modulation or strategic influence: object categorization (Casaponsa et al, 2024), object category learning (Maier & Abdel Rahman, 2024), embodiment of perceived power (Wei et al, 2024), and motion transitivity (Xue & Williams, 2024). Finally, and perhaps achieving the greatest level of detachment possible from language representations, the study by Dobler et al (2024) offers insights into how abstract concepts can be grounded in a biologically constrained model of semantic memory.…”

Transient and Long‐Term Linguistic Influences on Visual Perception: Shifting Brain Dynamics With Memory Consolidation

“…Online Versus Offline Effects on Perception Many studies have found CP for categories such as colors, familiar objects, or familiar animals, also comparing speakers of different native languages or bilinguals (Athanasopoulos et al, 2010;Boutonnet et al, 2013;Casaponsa et al, 2022;Drivonikou et al, 2007;Gilbert et al, 2006;Regier & Kay, 2009;Thierry et al, 2009;Witzel & Gegenfurtner, 2016). These categories are based on extensive prior experience, and language may have shaped preexisting mental representations of these categories in long-term memory, which makes it difficult to distinguish between online and offline effects.…”