Experiment 4: Read

Description: The experiment is intended for 4 different language groups: English monolinguals, Russian monolinguals, Russian bilinguals (L2 – English at a high level) that performed the test in English, and Russian bilinguals (L2 – English at a high level) that performed the test in Russian. There are 2 parts: Scaling Task and Sorting Task. In the Scaling Task subjects are provided with 5 stories, 3 of which describe prototypical “envy situations” and 2 – prototypical “jealousy situations”. The task is to scale the appropriateness of words “envious” and “jealous” (or their translation equivalents “zaviduet”/”revnuet”) to describe each of the stories. In the second part of the experiment subjects read 5 triads of sentences. The triads are the following: EEJ, JJE, EEN, JJN, EJN (where “E” stands for a sentence, describing an “envy situation", “J” – “jealousy” situation, “N” - general negative situation). In each triad subjects choose 2 sentences that, in their opinion, go together according to their similarity.

Discussion: In the first task we find out that there is an asymmetry in the way that English and Russian speakers (regardless of their L1) label emotions of envy and jealousy: In Russian language the scope of terms “revnuet” and “zaviduet” overlaps only marginally. In English, both words could be applied to “envy situations” .

In the second part of the experiment we found out that all four groups of speakers can reliably differentiate between envy and jealous situations. So it is not that in some cases English speakers use both terms interchangeably because they don’t see the difference. However, they are much more likely to group envy and jealous situations together, when opposed to a 3rd control negative situation. This shows that English speakers (both native and non-native) see the concepts as distinct though similar. The only group with very different results was the group of Russian monolinguals. So we may say that exposure to a foreign language with different conceptual structure leads to a conceptual shift.

In the previous experiments we investigated how L1 and L2 are interconnected, and how they access a conceptual level. However, as seen from the last experiment, bilingual conceptual level itself is not necessarily homogeneous. This finding is best of all captured in the De Groot’s Distributed Feature Model.

Figure 3: The Distributed Feature Model

There are 3 ways in which concepts of L1 and L2 may relate to each other:
1) conceptual equivalence:
Table (Eng.) = Tisch (Germ.). / Vader (Dutch) = Father (Eng.)
Since the concepts are fully equivalent there is no conceptual interference.
2) conceptual partial equivalence:
Jealous (Eng.) – Zaviduet (Russ.)
Asymmetry of conceptual representations is a possible source of interference and, as a result, non-native like performance.
3) conceptual non-equivalence:
Schadenfreude (Germ.) - Emmm… when someone is enjoying other’s misfortune (there is no direct equivalent in English)
Since there are no conflicting conceptual representations, this is not a source of language interference.

Read Overall Conclusion »