Résumé
Over the past two centuries, agricultural production methods have been pushed towards monospecific systems intensive in chemical inputs, with negative impacts on the environment and climate. A global awareness of the need to change the production patterns is gradually emerging (Sanchez et al., 2022). The aim is to identify synergetic agro-ecological solutions that are both socio-economically viable and environmentally friendly, as quickly as possible. Given the complexity of these issues, it is essential to recognise the value and relevance of informal/local knowledge and to build agroecological innovations around it (Sumane et al., 2018). Hybridization of knowledge is one way of achieving this; it is defined as the connection of knowledge from different sources with a view to action (Loyce et al., 2021). Co-designing agricultural practices may greatly benefit from this action-oriented knowledge generation. Our objective was to explore/experiment knowledge hybridization through in-plot exchanges on the agronomic functioning of cocoa agroforestry systems (AFS) and associated practices, in order to assess its hybridizing potential.We developed an innovative methodology for analyzing the hybridization of knowledge that took place between five participants, including three cocoa farmers, an ethnobotanist and an agronomist. As our main objective was to ensure farmers benefited from the project, the evaluation methodology was focused on what cocoa AFS knowledge they gained from this experience. The methodology consisted of three different phases: (P1) an initial phase, during which we used mind maps to take stock of each participant's individual knowledge, (P2) a phase of exchange between the different participants, on each of the three farmers' AFS, to review and compare the individual knowledge, whether similar or divergent, identified in P1, and (P3) an individual debriefing to obtain feedback from each participant on their perception of the approach and to assess the hybridization, i.e. the knowledge learned or transformed between P1 and P2. The work was carried out in Côte d'Ivoire in the plots of three farmers from three different cocoa-growing localities but within the same phytogeographic area, so as to increase the likelihood that their cocoa-growing knowledge would be varied and complementary.The results show that the participants learned new knowledge, notably about the benefits or drawbacks of certain shade trees on the provision of services and disservices. For example, one farmer learned that Ricinodendron heudoletii (Akpi) would supply water to the cocoa tree from the soil in the dry season via root system, while another understood that the infestation of borers on these cocoa trees was favored in the proximit of Triplochiton scleroxylon (Samba) as it harbored and allowed the proliferation of these borers. Some farmers were also able to learn about planting and pruning techniques, biofertilization and treatment with biopesticides. It is particularly interesting to note that information exchanged by farmers percolates far better than information exchanged between researchers and farmers. Farmers were particularly interested to hear from first hand experiences, and keen to imagine new trials to test their assumptions. This learning process remains quite successful, with around 46% (minimum 26% and maximum 71%) of topics for which farmers learned something between P1 and P3. On the other hand, participants' evaluation of the methodology indicates that they are very appreciative of the approach, and would like to see more actions of this type implemented by their agricultural cooperatives or by national technical advisory institutes.The hybridization process has undoubtedly resulted in the transmission of information in addition to the knowledge already acquired but not explored in phase P1. Unfortunately, our evaluation method, perhaps too restrictive, does not capture this hybridization. It would also be interesting to be able to dissect plot exchanges using other analysis methods from the language sciences and didactics, which would probably enable us to better understand why knowledge exchanged by peers is better transmitted. This raises the question of how we researchers can become part of an exchange between farmers to achieve this hybridization. Finally, one of the limitations of our exploratory approach is that we have only looked at verbal exchanges in the plot. It would therefore seem appropriate to use the same methodology to try and evaluate other types of knowledge transmission tools, such as written, audiovisual and other media commonly used to hybridize knowledge.