Theories about hominization, taking into consideration dietary adaptations and brain size increase, are discussed after a review of major data on primate feeding behavior, including a summary of the author's papers. Field work on primate published since the seventies shows a persistence of ancient morphological, physiological and behavioral adaptations in the present food responses. For instance, distribution of insects determines the size of individual territories of prosimians; however the ‘primitive' overlap of male and female territories also corresponds to spatial utilization of food resources by males and females in a group of chimpanzees. In any case, insects and other invertebrates, which are still providing an important protein input in the diet of forest people, remain attractive gustatory rewards and a necessary supplement for the frugivorous diets of most primates. The relationship between group size and structure, territorial behavior, and spatial distribution of food resources is exemplified by two species of langurs (Presbytis entellus and P. senex) separated in food niches although morphologically very closed. The ‘langur model' shows how feeding strategies can evolve, with the utilization of common plant species at low energy cost, or that of most scattered food plants providing high energy reward at a higher energy cost. Coevolution of animals and plants allowing such adaptive responses in terms of dietary specialization can be understood in terms of gustatory responses associated with potential toxicity and/or nutrient. The ‘gusto-facial reflex' of the newborn primate, associated with sweet or bitter tastes, leads to parental responses in humans, improving the beneficent feedback of sugar intake. However, for all primate species, the allometric relationship between sugar sensitivity and body size is the common rule, and deviations from a global tendency are associated with specialized diets. For bitter or astringent compounds such as alkaloids and tannins, the large variation in taste sensitivity of the primate species results from adaptive responses to environmental composition. Examples are given from studies in Gaboon and from the observations in progress in Madagascar. Such compounds may also have a beneficent action, although the term ‘medicine' does not appear as adapted in this context. Salts may also have ambivalent activities, in terms of potential toxicity and/or necessary food supplement. In parallel with the sensory adaptive responses to various environments, primate digestive morphology is the most obvious adaptation. The allometric relationship between digestive tract and body size has shown clear-cut categories corresponding to major dietary patterns of primates. This relationship locates humans among the frugivorous species (with diets including also seeds, some leaves and eventually some insects and meat as supplements). The issue about the increase of brain size in the genus Homo is discussed in this context, implying a high energy density of the diet. The likely early use of the fire (as early as 1.7 million years) that has been proposed by Wrangham et al., and the utilization of cooked tubers of wild yams (Dioscorea spp.) appears as the most likely hypothesis. Preparation and consumption of elaborated cooked foods would have been a major factor allowing brain size increase in the genus Homo, with a feedback facilitation of relationships and skills in the social group., La confrontation des théories actuelles de l'hominisation avec un ensemble de données sur le comportement alimentaire des primates non humains (publiées au cours des trois dernières décennies) permet de préciser les influences respectives des facteurs biologiques et culturels dans le déterminisme des choix alimentaires. En ce qui concerne les bases morphologiques des adaptations au régime alimentaire, cette revue (qui inclue et résume la plus grande partie des travaux de l'auteur), met en évidence un fonds commun à tous les primates que l'on retrouve chez l'humain. Elle permet ainsi de situer le point de bifurcation à partir duquel les influences de la transmission culturelle des préférences alimentaires dans les groupes d'hominidés deviennent prédominantes par rapport aux tendances que la socio-écologie a nécessairement modelées en fonction des environnements respectifs des différentes espèces de primates. Les variations observées de la perception gustative des différentes espèces constituent des clés d'interprétation dont l'étude serait encore à approfondir. En retour de ces considérations sur l'évolution des régimes alimentaires des primates, l'émergence des espèces d'hominidés à gros cerveau peut être expliquée selon un processus qui remet en cause les facteurs bioénergétiques et les variations des perceptions gustatives.