Multilingual word ratings


Many bilinguals report that processing emotion (happy) and emotion-laden (death) words in a second language (L2) doesn't have the same emotional impact as processing the same words in their native language. Several explanations why this might be the case can be considered. It could be that the meaning of L2 is only incompletely acquired (i.e. Perfetti's lexical quality hypothesis). This is not unreasonable because many emotion words are abstract, and the meaning of abstract words is relational and contextually defined, making them difficult to process. Others have proposed an embodiment explanation (e.g. the Affective Embodiment Account, Kousta et al., 2011). According to this explanation, emotional terms are difficult in L2 because non-native speakers acquire these words in a classroom setting rather than an environment where these emotions are experienced.  In addition, it could be a combination of factors that explain why emotion terms are represented differently in L2.

Research Questions / Hypotheses

The goal of this study was to collection affective norms for both valence (good/positive vs bad/negative) and arousal (high vs low arousal) for a large set of English words and their Mandarin translations. We also collected data about language background and history, together with an English vocabulary test.

We expected that ratings would be different between native and non-native speakers. In line with previous cross-cultural work, we expected high arousing words to be perceived less positive for Mandarin speakers. According to the lexical quality hypothesis we expected that more proficient participants would have ratings that resemble native speakers.


A total of 165 native English speakers and 182 native Mandarin speakers judged English words. There were also 76 native Mandarin speakers who judged the Mandarin translation equivalents of the English words.


The study consisted of several questionnaires to measure language experience and history, together with a normed vocabulary test (LexTale) in which participants had to decide whether a word is an existing English word or not. Participants completed both valence and arousal ratings for a subset of nearly 1000 words using a 9-point bipolar rating scale. In addition, we also used a control by having a group of Mandarin speakers judging English words and a separate Mandarin group judging Mandarin translations of these words.


Data collection is still ongoing as we aim to have a balanced sample of male and female raters and the sample size for male Mandarin speakers is not sufficient. Balancing the norms for gender is important as previous work suggests systematic variability as  a function of gender.

So far our results confirm that valence rating are highly reliable, but arousal ratings require larger samples to be achieve comparable levels of reliability.  Since we are still collecting data it is too early to draw conclusions about systematic differences between native and non-native speakers or between highly and moderatly proficient speakers.


Lexical norms for bilinguals are often limited to Indo-European languages, whereas similar norms for  widely spoken language like Mandarin are lacking. Apart from theoretical contributions regarding the role of emotion processing in L2, we expect these norms to be used in other psycholinguistic studies that aim to manipulate or control the emotionality of stimuli words in lexical processing tasks.