From a semantic point of view concrete nouns can be divided into different categories or domains such as living - non-living as well as into different lexical hierarchies (e.g. dog - beagle). Psycholinguistic and aphasiological case studies as well as neurophysiological investigations have shown that distinct semantic categories can be selectively damaged or show advantages in processing. However, these mental lexicon structures are still sketchy with regard to psycho- and neurolinguistics. It is still unknown if semantic categories and lexical hierarchies are just a theoretical construct or if they have a cognitive reality. Neurolinguistic approaches vary in their assumptions as to whether mental lexicon structures are feature-, evolution- or structure-based representations.
30 subjects (15m/15f) from 19 - 33 years (av. 23,8 ± 4, right handed) participated in the experiment. We used Electrocap with 32 scalp electrodes (10/20-system) and 4 eye electrodes. We conducted t-tests and ANOVA.
We presented 240 spoken words from 12 domains (= 6 pairs, 20 stimuli per domain) as well as 160 colour pictures from 8 categories (= only 4 pairs, because some categories can not be depicted, 20 stimuli per category). The semantic domains were: living (e.g. bee, tongue) vs. non-living (e.g. cover, shelf), edible (e.g. apple, ham) vs. inedible (e.g. car, garden), functional (e.g. anchor, scrubber) vs. sensory (e.g. bell, cactus), man-made (e.g. cup, paper) vs. natural (e.g. oat, fir), countable (e.g. coin, soap) vs. not countable (e.g. iron, hail) and hyperonym (e.g. cutlery, insect) vs. hyponym (e.g. nippers, fork).
Significant results in the auditory condition can be found in the distinction between two different lexical hierarchies. Listening to two lexical hierarchies can evoke strong differences in their amplitude course. Auditorily presented hyperonyms (superordination) elicit a higher and prolonged negativity in contrast to hyponyms (subordination) in almost all scalp electrodes.
There where no significant results in the following category pairs: living vs. non-living, edible vs. inedible, functional vs. sensory, man-made vs. natural as well as countable vs. not countable.
We found significant results in all categories in the visual condition. The amplitude courses of different categories have all one thing in common: the presentation of "sensory quality" objects (which are represented by taste, scent, tactile sense) elicits higher negativity around 200 ms post word-onset and lower negativity around 550 ms post word-onset in comparison to more functionally represented objects (e.g. tools, vehicles, machines).
The results in the acoustic condition support the assumption, that these hierarchies are processed differently on a neural level. An assumption of this is the fact, that hyperonyms contain only little sensory/perceptual information and are less specific in their semantic concepts in contrast to hyponyms, which are more concrete and thus have more specific semantic concepts and which contain a lot of sensory/perceptual information.
The differences between the amplitude courses of the contrasted categories in the visual condition are similar. The fact that the amplitude courses for the "sensory quality" categories are all similar in contrast to those of more functionally represented objects leads to the assumption that sensory/perceptual knowledge is processed differently to functional knowledge about objects on a neural level.
The results of the auditory and visual presentation of different hierarchies and categories show that conceptual knowledge may be organised through different conceptual structures. The fact that the differences beween the semantic categories in the visual condition are much stronger than in the auditory condition can be explained by the picture-superiority-effect.