TACO: A Turkish database for abstract concepts

Abstract

The organization of abstract concepts reflects different dimensions, grounded in the brain regions coding for the corresponding experience. Normative measures of linguistic stimuli offer noteworthy insights into the organization of conceptual knowledge, but studies differ in the dimensions and classes of concepts considered. Additionally, most of the available information has been collected in English, without considering possible linguistic and cultural differences. Here, we aimed to create a comprehensive Turkish database for abstract concepts (TACO), including rarely investigated classes such as political concepts. We included 503 words-78 concrete (fruits, animals, tools) and 425 abstract (emotions, social, mental states, theoretical, quantity, space, political)-rated by 134 Turkish speakers for familiarity, imageability, age of acquisition, valence, arousal, quantity, space, theoretical, social, mental state, and political dimensions. We calculated dominance and exclusivity, indicating the dimension receiving the highest mean score for each word, and the position of the word along the unidimensional–multidimensional continuum, respectively. A principal component analysis (PCA) was conducted on the semantic dimensions. The results showed that mental state was the dominant dimension for most concepts. Moderate to low levels of exclusivity indicated that the concepts were multidimensional. PCA revealed three components: Component 1 captured the juxtaposition between social/mental state and magnitude polarities, Component 2 highlighted affective components, and Component 3 grouped together political and theoretical dimensions. The introduction of political concepts provided insights into the multidimensional nature of this unexplored class, closely intertwined with the theoretical dimension. TACO constitutes the first comprehensive Turkish database covering several abstract dimensions, paving the way for cross-linguistic and cross-cultural studies of semantic representations.

Appendices

Appendices

Appendix 1 Alignment between the a priori classification and data driven results

We explored the extent to which the a priori classification into abstract and concrete domains aligned with the imageability (IMG) ratings. Using the median split we divided the dataset into concepts with low IMG (mean rating < 3.6, n = 248 concepts) and high IMG (mean rating ≥ 3.6, n = 255 concepts). A good alignment was reported, with 97.5% of the a priori classified concrete concepts rated as with high IMG, and only 2.5% with low IMG. The majority of the a priori classified abstract concepts (57.9%) were rated with low IMG, while 42.1% with high IMG.

We then explored the extent to which the a priori classification into categories aligned to the classification derived from the dominance analysis. We calculated the percentage of concepts for each abstract a priori category receiving as dominant one of the 6 semantic dimensions empirically collected, e.g., mental state, social, political, quantity, space, and theoretical. Overall, we found a good alignment (see Table 5 below). We excluded the concrete a priori categories from the current analysis of alignment.

Table 5 Alignment between the a priori classification and the classification derived from the dominance analysis, expressed in percentage

In detail, for the a priori classified theoretical, political, and mental state concepts, there was a very good agreement with the relative dominant dimension (i.e., more than 80%). Less alignment emerged for the concepts a priori defined as belonging to the social and space-quantity categories, for which only 50.43% and 50% (space in 20.24% and quantity in 29.76%) of cases, respectively, were classified in the respective dimensions. A particular case was represented by the a priori classified emotional concepts, since emotion was not present among the rated dimensions. For 91.67% of the a priori classified emotional concepts, mental state dimension was rated as dominant. This result can be explained by the definition of mental state used for the rating procedure, that can easily encompass affective and emotional states, namely a ‘set of intentions, moods, thoughts, purposes, cognitive evaluations that an individual experiences for himself or that he recognizes and attributes to other people’.

Appendix 2 Distribution of abstract and concrete categories across PCA components

We explored the distribution of abstract and concrete categories across the components of our Principal Component Analysis (PCA).

The majority (more than 89%) of concepts a priori classified as emotion, political, social, and mental state showed positive scores in Component 1, indicating that they were rated as connected to mental state and social dimensions, and as weakly connected to space and quantity. The opposite pattern emerged instead for the a priori classified space and quantity concepts. The concepts defined a priori as theoretical were crosswise distributed in Component 1, with concepts as felsefe (philosophy) receiving high component scores, i.e. being judged as associated to social-introspection, and concepts as logaritma (logarithm) receiving low component scores, i.e. being judged as associated to magnitude. Additionally, all concrete concepts displayed negative values in this component, meaning that, as expected, they were judged as weakly connected to the social-introspection dimension, and as highly connected to space and quantity. For instance, among concrete concepts the word cetvel (ruler) received the most negative component score.

Both concrete and abstract concepts were crosswise distributed in Component 2. For example, words as tabanca (gun) received high component scores, i.e. being rated as negatively-valenced and associated with high arousal, whereas words as tavşan (rabbit) received low component scores, i.e. being rated as positively-valenced and associated with low arousal.

Positive scores in Component 3 were found for 75% and 95.96% of the a priori classified theoretical and political concepts, while in the a priori category of emotion only 9.52% of concepts had positive component scores. The distribution of the other abstract a priori categories, i.e. social, mental state, space and quantity, across Component 3 was mixed. For example, the social category included words as hiyerarşi (hierarchy), being judged as highly connected to the political and the theoretical dimensions, but also words as alay (derision), being judged as only weakly connected to these dimensions. Finally, the majority of concrete concepts received negative values in Component 3, meaning that they were rated as weakly connected to both political and theoretical dimensions.