This study investigates how multilinguals compared with monolinguals categorise landscape. The aim is to contribute to research already done with fresh data in a new setting. The thesis suggests that speakers with lower proficiency in a language will tend to choose vocabulary based on frequency and not semantic knowledge. The study was done through a quantitative and qualitative questionnaire targeting French, English, and Swiss-German speakers. Methods used were among others word-to-referent mapping and priming to gather useful vocabulary. The participants were to look at 30 images per language and choose among three words which one best fit the images. All images portraying different features of landscapes. What could be found is that 20% of the time the multilingual had a tendency to behave differently to the monolingual baseline. Data from this study shows that the Indo-European languages in this research seem to have more similar constraints on the semantic domains than differences. 80% of the time all speakers, monolingual and multilingual, agreed on what word best fit the image. Qualitative and quantitative data support this idea. The frequency theory proved to be true in the 20% of cases where the trilingual group deviated from the answers of the other groups.