INTRA-CULTURAL CONTINUITY AND DIVERSITY OF SEMANTIC CATEGORY NORMS

Introduction

Culture and conceptual behavior are inseparable [6]. The frequency of concepts generated in specific semantic categories (like “A Toy” or “A Furniture”) can differ dramatically between different languages. Thus, a frequency of words can be used as a key indicator of the specific features of cultures [12].

Cultural differences can be caused by many factors, such as differences in social, historical, economic, educational systems and habitat, the degree of familiarity of the population with the category [7]. It has been shown that some categorization phenomena can be quite sensitive not only to language and to culture aspect but to experiential factors as well [5], [13], [11]. Nevertheless, we can’t suggest that only environmental factors will explain category representation as thanks to culture and language people can build the representation of things they did not interact with [10]. It can be suggested, the culture, which is dominated in the country, can outweigh environmental factors, which can vary in different regions. However, it has been shown in many studies that there is a crucial intra-cultural diversity, which can be found even inside the country with a homogeneous culture, which can be explained by social and economic factors [8], [2].

Generation frequency database for 45 semantic categories was collected for the Russian language [3]. Category norms collected in Moscow were shown to be reliable [3]. Nevertheless before making inferences and generalizing generation frequency norms collected in Moscow to the Russian language and the whole country, the geographical stability of these results was examined. Moscow, Irkutsk, and Yekaterinburg regions were chosen for this aim. As correlations between the three regions were very strong, the geographical stability of generation frequency norms for the Russian language was suggested [4]. Such result supports the idea that cultural representations maintain their stability while reaching a cultural level of distribution [10].

Nevertheless, previous work was aimed only to prove geographical stability and further analyses needs to be continued in order to study the intra-cultural homogeneity of category norms and regional specificity of concepts with more sensitive statistic methods. Hellinger Affinity, which is used in the current study, as a measure of distance between frequency distributions, is much more sensitive to differences than correlation. Furthermore, a separate comparison of frequency for each concept in a category was conducted with help of Chi-square test with Yates correction in the study.

Thus, the aim of this study was to measure Hellinger Affinity distances in order to evaluate a degree of intra-cultural homogeneity. and to compare differences in generation frequencies between three cities in order to understand the degree of regional specificity of concepts.

The following suggestions can be made. Between-regional accordance of generation frequency norms will be strong showing stability of representations on a cultural level. Nevertheless, the strength of differences can be related to geographical distance. As cities are closer to each other, fewer differences can be observed.

Method

Participants. One hundred sixty-two students of different universities of Moscow aged 18-26 years participated in the study as volunteers (126 females and 36 males, M=19.19, SD=1.52). One hundred sixty-two students from Yekaterinburg aged 18-24 years (126 females and 36 males, M=19.56, SD=1.18) and one hundred sixty-two students from Irkutsk aged 18-28 years (132 females and 30 males, M=19.31, SD=1.82) participated in this study as well. There was no significant difference in proportion of male and female participants in samples taken from three different regions (Pearson Chi-square=0.685, df=1, ps=0.408, Pearson Chi-square=0.000, df=1, p=1). Participants involved in this study in each region represented the same generation and the same social group. All of the participants were native Russian speakers.

Procedure. The procedure used to gather the Russian category norms was similar to the procedure of Battig and Montague (1969). Participants were asked to write down in the notebook as many items included in each category as they can in 30 sec. For example, if they were given category “birds” they were supposed to write names of different birds like a sparrow, a nightingale etc. The full version of the instruction can be found in the paper by Battig and Montague of 1969 [1]. Overall generation frequency was calculated for each word. The category set for this study consisted of 45 different semantic categories [4]. For each word, overall generation frequencies were compared between regions with help of two-tailed Pearson Chi-square with Yates` continuity correction. The number of significant differences was compared between regions using Chi-square-test as well. Hellinger Affinity coefficients as a measure of distance between two frequency distributions from different regions were measured for each category. The value of Hellinger Affinity coefficients ranges from 0 to 1 and the 1 means that two distributions are identical. Data were analyzed with help of Microsoft Office Excel and IBM SPSS Statistics 23.

Results and Discussion

Hellinger Affinity coefficients were measured between frequency distributions of three regions (table 1). The level of Hellinger Affinity was strong proving the intra-cultural stability of such kind of data. The similar level of intra-cultural affinity was received in other studies [14], [9].

Table 1 – Hellinger Affinity coefficients as a measure of distance between frequency distributions

Hellinger Affinity scores between frequency distributions of Moscow and Yekaterinburg samples were stronger than the Hellinger Affinity scores between frequency distributions of Moscow and Irkutsk samples (Z=-2.261, p=0.024). The presence of significant differences in levels of Hellinger Affinity shows some degree of slight regional diversity. Differences in the strength of Hellinger Affinity between other two pairs of regions were insignificant (p˃1).

Chi-square comparison with Yates` correction showed that there were no significant differences between three cities for the most concepts. Differences in generation frequencies were significant (p˂0.05) only for 3.98 -5.24 % of concepts (table 2). Just a small number of concepts showed very significant differences with p˂0.01 (1.27-2.41%). This finding once again supports the suggestion that norms of generation frequency for the Russian language are geographically homogeneous.

Table 2 – The number of words with significantly different frequencies between regions

M: Moscow, I: Irkutsk, Y: Yekaterinburg

Such small percent of differences can be explained by random factors and ignored. Nonetheless, some regional specificity can be revealed in such limited number of concepts. That is why the number of concepts with significantly different generation frequency in samples of three regions (two-tailed p<0.01) compared using Chi-square without Yates` correction.

There were significantly more concepts, which differed in generation frequency between Moscow and Irkutsk than between Moscow and Yekaterinburg (Pearson Chi-square=19.835, df=1, two-tailed p˂0.0001). There were no significant differences in the number of concepts between Moscow-Yekaterinburg and Irkutsk-Yekaterinburg (Pearson Chi-square=1.722, df=1, one-tailed p=0.1894). There were more concepts, which had significantly different frequencies between Moscow and Irkutsk than between Yekaterinburg and Irkutsk (Pearson Chi-square=9.963, df=1, one-tailed p˂0.01).

Thus, it can be suggested that strength of similarities related to geographical distance. Stronger similarities are observed between Moscow and Yekaterinburg, which are closer to each other than between Moscow and Irkutsk.

Analyses of words with strongest significant differences in frequency between regions showed that in some cases intra-cultural diversity could be explained by ecological, environmental and experiential factors. For example in Irkutsk sample word omul (fish with lives in Baikal lake which is close to Irkutsk) was named 90 times while in other samples it was named twice (Pearson Chi-square with Yates` correction= 114.897, df=1, two-tailed p˂0.0001 in comparison to Moscow and Yekaterinburg). In Moscow sample in category “vehicles” metro was the frequent answer, while in Irkutsk it was not as there is no subway (Pearson Chi-square with Yates` correction = 67.258, df=1, two-tailed p˂0.0001). The frequency of word metro was significantly lower in Yekaterinburg sample as this type of vehicle is not such popular in this city as it is in Moscow (Pearson Chi-square with Yates` correction =25.372 df=1, two-tailed p˂0.0001).

The intra-cultural affinity for data is shown to be strong. Most of the concepts (more than 97.5%) did not differ in frequency between regions. Even though such rare significant differences in frequency can be explained by random factors, some cross-regional diversity for a few concepts was revealed and should be taken into account. Such differences can be explained by geographical, socio-economical, environmental and experiential factors.