Experimental conditions
To better understand how subjects could associate images with flavors, a little experiment with ten people was carried out. Be aware that given the nature of the sample, the results should be interpreted with caution.
The task was to associate with 4 zakouskis, images taken from a set of ten photos of architectural works presented in the south of France during the 2016 festival of architectures (FAV) in Montpellier and La Grande Motte. The 10 photos of “architectural elements” are given above.
The 4 zakouskis were : a toast with Maredsous cheese, salmon and chives; a tabbouleh with carrot juice, green peas and shrimps; a roulade of chicken breast, stuffed with speculoos , chicons and finally a macaroon with foie gras and blackberry coulis.
The instruction was to choose the 4 photos that best married the 4 zakouskis. Moreover, I had stressed the importance of the sequence of associations. It was important to remain consistent in the proposed links.
Main results
The results were analyzed in depth and can be summarized as follows:
- High diversity of choices: there seems to be no consensus on the proposed associations. Even when the proximity between images is taken into account, subjects do not proceed in a similar way to associate images with zakouskis.
- No strong preferences for a limited set of photos : there are at least 5 different pictures for each zakouski. All photos were chosen at least once and photo 3 was 8 times. Note however, and this will be commented on later, that 30% of the subjects chose for the toast the picture 7, 40% associated with the tabbouleh the photo 3 and 40% associated with the « roulade » the picture 10.
- In subsequent analysis, I tried to characterize the logic used by the subjects : do they map images which are similar to tastes which are similar ? To answer this question, I had to define a distance on the zakouskis based on evaluation grids used by experts (oenologists, …). This made it possible to identify an optimal choice by using my mapping device and to study the logic of the choices made by the subjects. The subjects did not seem to proceed like the mapping device ; « close » tastes were not necessarily associated to close images. This conclusion should be taken with care as the zakouskis were fairly different and the photos as well.
- On the other hand, it is not impossible that some choices were inspired by visual similarities between the photos and the zakouskis: the toast with a flat and rectangular shape and the photo 7 (chosen by 30% of the subjects) or the « roulade » and photo 10 (chosen by 40% of the subjects).