| Keywords [eng] |
left-digit bias, left-digit effect, numeral processing, round numerals, modified numerals, approximation, degree semantics, kairiojo skaitmens nuokrypis, skaitvardžių apdorojimas, apvalūs skaitvardžiai, apytikslė reikšmė, laipsnių semantika |
| Abstract [eng] |
Originally described as a somewhat puzzling aspect of erroneous price judgements (e.g., judging $4.99 as closer to $4.00 than to $5.00), left-digit bias has since been observed across a wide range of domains, such as medicine, education, and competitive sports. Individually, these observations have all been interpreted as consequences of the imprecise mental mapping between abstract magnitudes (i.e., •••••) and their symbolic representations (i.e., 5), invariant across all aforementioned domains, dimensions, and scales. However, much less work has been done in order to reconcile this psychological perspective with the principles of pragmatics, governing the context-specific constraints of numeral meaning. In an effort to bridge this gap, the current thesis aims to establish how left-digit bias, valence, and selected linguistic features contribute to speakers’ numeral interpretations. Using a combination of experimental and corpus data, aspects of left-digit bias are examined across various scales, sentential contexts, and languages. Based on data from 116 participants, the experimental portion reveals the degree of bias exhibited by English speakers across 30 different measurement scales and sentential contexts, as well as its lack of interaction with attribute-level affective valence. A context-sensitive model of numeral processing is proposed on the basis of these findings. The main assumption of the model is corroborated by corpus data from three different languages. However, further research is needed to establish the sensitivity of the bias to additional linguistic features. The implications of these findings for the relevant subfields of psychology and linguistics are discussed. |
