Taste perception and food choices in capuchin monkeys and human children

Summary

Despite more than 40 million years of independent evolution, capuchin monkeys and human children share several features that make a comparison in the domain of feeding behaviour interesting. As with humans, capuchin monkeys have a long life span and an extended infancy period; moreover, they are omnivorous and food neophobic. In both species, taste provides an immediate and powerful feedback when selecting foods. In humans, acceptance and rejection responses are evident beginning in early infancy, before experiencing any consequences from the ingestion of sweet or bitter substances. Similarly, capuchins initially prefer novel foods with a high sugar content that is readily perceived through taste. However, after repeated encounters with these foods, capuchins change their preferences, responding to the feedback coming from the foods' energy content, in order to maximize the net gain of energy. Also in children, positive consequences of the ingestion of a food can be associated with the flavour of that food and can increase its consumption. Preschool children learn to prefer food with a high caloric content over food with a low caloric content and use different flavours as immediate cues to distinguish foods. Another factor influencing the consumption of a novel food is how often it is encountered. For capuchins, a food remains unfamiliar only for the first few encounters. Similarly, children’s neophobic response decreases with repeated exposures to novel foods. Furthermore, in both species social influences may help to overcome food neophobia and to accelerate the acceptance of novel foods into the diet. In conclusion, we argue that capuchin monkeys provide a good model for investigating the factors affecting the acquisition of diet in human children.

INTRODUCTION

Non-human primates have repeatedly been shown to be useful models for human taste perception, broadening understanding of how diet selection and food intake are controlled in humans (Drewnowski, 1997). Whereas environmental factors — such as beliefs and attitudes concerning weight and dieting — modify food preferences of human adults, in very young children, and non-human primates, many of the considerations that influence adult food acceptance patterns do not apply. In children, only taste and familiarity account for food preferences (Birch, 1979a, 1979b; Drewnowski, 1997). In this paper we will present data on the feeding behavior of capuchin monkeys and children that show how despite more than 40 million years of independent evolution, they share several features that make comparing their feeding behavior extremely interesting.

Like humans, tufted capuchin monkeys (Cebus apella) are omnivorous. This species, which is widely distributed in South America, has a very broad diet consisting mainly of plant food, especially fruits, and a variety of invertebrates (Kinzey, 1997). Capuchins have been described as a successful genus (Fragaszy, Robinson and Visalberghi, 1990), whose ecology and diet are extremely adaptable (Kinzey, 1997). Moreover, capuchins have a long life span (the record is 53 years, Visalberghi and Anderson, 1999), long infancy and juvenile periods (Fragaszy and Bard, 1997), large brains relative to body size (Bauchot, 1982; Stephan, Baron and Fraham, 1988; Marino, 1995), and are born with brains that are relatively undeveloped and neurally immature (Fragaszy and Bard, 1997; Ross, 1991).

In this contribution, we aim (1) to provide a description of several factors that might affect food choice and ingestive behavior in omnivorous species, (2) to review the most relevant studies on this topic conducted with capuchin monkeys and human children and discussing the data, whenever possible, using a comparative perspective.

FOOD NEOPHOBIA: AN ADAPTIVE ANSWER TO THE OMNIVORE’S DILEMMA

The success of an omnivorous species depends on its propensity to explore and sample novel foods and to include them in the diet, as well as on its caution toward them, in order to detect and to avoid the risk of ingesting poisonous substances (Freeland and Janzen, 1974; Glander, 1982; Milton, 1993). Paul Rozin labeled this conflict as the “omnivore’s dilemma” (Rozin, 1977). A partial solution to the omnivore’s dilemma is food neophobia, i.e. the hesitancy to eat novel foods (Barnett, 1963). However, in the long run food neophobia is maladaptive, because omnivores increase chances of survival through variety in the diet. Therefore caution towards new, potentially toxic, substances should be balanced with this need for variety (Birch, 1998; Scott, 2001). In fact, food neophobia is a behavioral response that prevents the ingestion of large quantities of a novel food, and thus minimizes the risk of being poisoned (Glander, 1982; Freeland and Janzen, 1974). At the same time, the ingestion of a little amount of a novel food is sufficient to induce post-ingestive consequences (see below), leading the animal to avoid or accept that food in its diet. Also animals capable of detoxifying and eliminating secondary compounds should treat novel foods with caution, in order to avoid serious damages to their detoxification systems (Freeland and Janzen, 1974).

Capuchin monkeys are neophobic both in captivity and in the wild. Visalberghi and Fragaszy (1995) presented 11 captive capuchin monkeys with 10 novel foods (never tasted by any of the subjects previously) and 2 familiar foods. Each subject was tested individually once with one food at a time. Acceptance of the 10 novel foods ranged from complete acceptance (all the 11 subjects tasted the food at least once) to unanimous rejection, and novel foods were consumed to a lesser extent than familiar ones. Similarly, when 10 novel foods were presented to a group of wild capuchin monkeys ranging from 25 to 30 individuals, results showed that capuchins approached and ate very little of the novel food; instead, familiar foods were completely consumed (Visalberghi, Janson and Agostini, 2003). However, capuchins living in a 43.000 ha ecological reserve (Parque Nacional de Brasilia, Brazil), which partially rely on foods left by the visitors, do not seem to show food neophobia. They have probably learned that humans leave behind foods that are safe to eat and are willing to exploit food sources left by humans in the area (Siemers, 2000; Visalberghi, pers. obs.).

Infant and youngster capuchins are less neophobic than adults. When Fragaszy, Visalberghi and Galloway (1997) investigated food neophobia in 11 infant capuchin monkeys (4.5–12 months), they found that infants did not treat novel foods with caution. Instead, infants picked up and ate novel foods more frequently than familiar ones. A similar trend has been shown for wild capuchins: youngsters were far more responsive than adults to the novel foods, and contacted, manipulated, explored, and ate the novel foods significantly more than adults (Visalberghi, Janson and Agostini, 2003). Since infants and juveniles are less efficient foragers than adults (Janson and van Schaik, 1993; Terborgh, 1983; van Schaik and van Noordwijk, 1989, Boinski and Fragaszy, 1989), their lower neophobia, compared to adults, may help them to avoid the risk of starvation.

Human children are cautious about novel foods, and their rejection of novel foods is a common cause of parental concern and frustration (Birch, 1983). As mammals, during the first months of life human children consume only milk. However, about halfway through the first year of life, an exclusive milk diet is no longer adequate to maintain growth and health, and the transition from milk to an omnivorous diet occurs (Birch, 1998). Experimental evidence showed that infants (4–7 months old) are less neophobic than older children (2–5 years old). Birch, Gunder, and Grimm-Thomas (1998) suggested that neophobia would not be a fully functional response during infancy, when food is provided by parents, whereas it becomes more important by early childhood, when children have begun to explore the environment and eat by themselves. Nevertheless, omnivorous species should overcome neophobia, because their success depends primarily on their ability to exploit new food sources during the acquisition of a varied diet and/or to face seasonal changes in food availability.

The lack of evidence of food neophobia in young capuchins (as also shown in young lemurs; see next chapter by L. Tarnaud) seems to contradict the heightened presence of this phenomenon in young children. More research is needed to determine whether methodological differences and/or developmental variations account for the differences in the neophobic behavior in children and young capuchins. In the latter species, neophobia has never been investigated in relation to weaning.

SWEET TASTES BETTER: THE ROLE OF TASTE PERCEPTION

Primates are able to assess food quality through the sensory information that comes from the foods they consume (Dominy, Lucas, Osorio and Yamashita, 2001). Senses, and taste in particular, enable primates to decide whether swallowing or rejecting a potential food, and there is increasing evidence that taste perception is an adaptive response to the animals’ need to assess food nutritional contents (Simmen and Hladik, 1998; Laska, Hernandez Salazar and Rodriguez Luna, 2000; Hladik, Pasquet and Simmen, 2002; Visalberghi, Sabbatini, Stammati and Addessi, 2003).

Substances high in calories, such as sugars—a very important energy source—are perceived as sweet by humans and are readily accepted by both humans and non-human primates (Glaser, 1993). On the contrary, the tendency to avoid bitter taste is associated with the presence of plant secondary compounds, such as alkaloids and glycosides, whose taste can function as a cue to inhibit their ingestion (Ueno, 2001). In fact, secondary compounds can exist in nature in concentrations far exceeding those required to cause death or severe physiological damage to almost any mammal that might eat them (Freeland and Janzen, 1974).

Human neonates and newborn non-human primates react to gustatory stimuli in stereotypical ways, showing differential orofacial motor reactions (gusto-facial reflex; Steiner and Glaser, 1984). They accept sugar solutions by sucking with relaxed facial expressions, and reject bitter solutions with a typically arched form of the mouth (Rosenstein and Oster, 1998). In one study, when sucrose was presented to infants in the form of a sucrose-infused gelatin nipple, the infants’ sucking response increased in frequency and intensity (Maone, Mattes, Bernbaum and Beauchamp, 1990). Sensitivity to bitterness is relatively unstudied in infants and young children. However, there is some evidence that bitter taste perception heightens in the first year of life (Kajiura, Cowart and Beauchamp, 1992). In this study, newborn infants (0–6 days old) did not readily distinguish various concentrations of urea, whereas older infants (14–180 days) tended to reject all concentrations of the substance. To our knowledge, the reaction of newborn capuchins to gustatory stimuli has not been investigated, but it is very likely that the gusto-facial reflex and sucking preferences for sucrose are present in this species as well.

INVESTIGATING FOOD PREFERENCES IN CAPUCHIN MONKEYS

Taste perception seems very relevant for food selection in tufted capuchin monkeys. Visalberghi, Sabbatini, Stammati and Addessi (2003) investigated the acquisition of preferences towards 7 novel foods presented to 26 individual capuchins. The experiment consisted of three phases: Pre-treatment, Treatment, and Post-treatment. In the Pre-treatment and Post-treatment, subjects were tested individually, while during Treatment one half (N = 13) of the subjects was tested individually (Individual condition) and the other half (N = 13) was tested together with 3–5 other group members (Social condition).

In the Pre-treatment, subjects were presented with all 21 possible binary combinations of the 7 foods using a two-alternative choice test (for similar studies on spider monkeys, Ateles geoffroyi, squirrel monkeys, Saimiri sciureus, and pigtail macaques, Macaca nemestrina, see Laska et al., 2000; Laska, 2001). Once the subject made a choice by taking one of the two foods, it could not take the other one. Treatment was aimed to familiarize the subjects with the foods presented in the Pre-treatment. During Treatment subjects received a 10-minute session per day for five consecutive days. In each session, the 7 foods were scattered on the entire floor of the cage. In the Individual condition, subjects had 8 pieces of each food. In the Social condition, the same quantity of food (8 pieces of each food) was available for each of the subjects (see below). In the Post-treatment, each subject was again presented individually with the 21 binary choices between the foods. The experimental procedure was the same as during the Pre-treatment.

Results of the Pre-treatment showed that, already after having eaten small portions of each novel food only a few times, capuchins markedly preferred certain foods over others. Their preferences were positively correlated with the glucose and fructose content of the foods. So, short-term feedback coming from taste seems to be important for establishing initial food preferences. This finding is in agreement with a previous observation showing that, among non-human primates, Cebus spp. have the highest sensitivity to sugars (Glaser, 1986). Since sweet taste is usually not associated with toxicity, it seems adaptive for an omnivorous species, such as Cebus apella, to rely on sugar content for choosing between novel foods and enlarging the diet without too much risk.

In the wild, palatability of plant foods may change over time in relation to the concentration of toxic secondary metabolites. Jones (1978) showed that the levels of cyanogenic glucosides in various plants differ depending on geographical locality. In addition, the concentration of toxic secondary metabolites (e.g. cyanogenic glucosides) changes in relation to plant age (McKey, 1975; Janzen, 1983), season (Ellis, Keymer and Jones, 1977a, b), and plant tissue type (Cooper and Johnson, 1984). In another study, Visalberghi and Addessi (2000) investigated capuchins’ behavioral response to a familiar food changing in palatability over time. Capuchin monkeys were presented with a familiar food whose palatability changed according to the experimental phase. In Phase 1 capuchins were individually presented with this familiar food. In Phase 2 they received the same familiar food with pepper added to it, making it unpalatable. In Phase 3 they received the same familiar palatable food of Phase 1. Five sessions were carried out in each phase. Capuchins readily adjusted to changes in flavor and palatability of a familiar food since the first session in which the change occurred. They reduced (Phase 2) and increased again (Phase 3) the amount of food eaten, and encounters with the food when unpalatable did not affect its consumption when palatable once again. Therefore, capuchins undoubtedly showed behavioral flexibility when facing negative as well as positive changes in the palatability of a familiar food.

INVESTIGATING FOOD PREFERENCES IN HUMAN CHILDREN

Research indicates that human infants’ early experiences with food flavor have decisive influences on their acceptance of food. Infants are first exposed to the flavor of food before they are ever born. Flavors are transferred from the mother to the fetus through amniotic fluid that is periodically ingested through the unborn child’s mouth and nose (Hauser, Chitayat, Berns, Braver, and Muhlbauer, 1985, Mennella, 1995). Moreover, newborns are exposed to flavors through breast milk if they are breastfed (Mennella and Beauchamp, 1991; Gerrish and Mennella, 2000). Children who are breast-fed are exposed to a variety of flavors, and there is evidence that babies prefer flavors they have previously experienced through breast milk (Mennella and Beauchamp, 1993; Mennella and Beauchamp, 1996; Mennella and Beauchamp, 1998). There is also evidence that breastfed children are more likely to accept pureed vegetables when first introduced to solids than formula-fed babies (Sullivan and Birch, 1994). In addition, the positive influence of breastfeeding on the reduction of picky eating is evident in children as old as 7 years of age (Galloway, Lee and Birch, 2003).

Recent findings suggest that infants who are exposed to a particular flavor before birth (via food flavors that pass through the placenta) or through breast milk, will prefer that flavor six months after birth more readily than infants never exposed to the flavor (Gerrish and Mennella, 2001). In this study, expectant mothers were assigned to one of three conditions: 1) they were asked to drink approximately one glass of carrot juice for 12 weeks in the last trimester of pregnancy, 2) they were asked to drink approximately one glass of carrot juice everyday for 12 weeks during after their babies were born, or 3) they were asked to drink approximately one glass of water 12 weeks before and after childbirth. Infants exposed to the carrot juice either prenatally or postnatally preferred cereal mixed with carrot juice over cereal mixed with water when solids were first introduced.

Birch (1979a) developed a procedure for directly assessing food preferences in young children. Children are individually presented with a tray of seven to nine foods, small samples of which are contained in separate transparent cups. The child tastes each item in a self-selected order. After tasting the food, the child is asked to place it in front of one of three cartoon faces that corresponds to the child’s affective response to the food: a smile (like), a frown (dislike), or a neutral expression (just okay). Then, the child is asked to rank the foods within each grouping (liked, disliked, and just okay) in order to obtain a complete rank order. This procedure is repeated for each of the three categories. Rank ordering within categories can be combined to comprise a complete rank order on the foods. Even though preferences may vary with time of day and through the course of a meal, for most of the three-year-olds and nearly all older children the preference data obtained in this way are reliable and predictive of the relative consumption (see also Birch and Sullivan, 1991).

In the absence of adult pressure, children usually eat more of the foods they like best (Birch, 1992). In the early years of life, food likes and dislikes are the primary determinants of food intake; children accept sweet foods more readily and are very likely to reject bitter or sour ones. In three- and four-year-old children, familiarity and sweetness are the two dimensions accounting for their food preferences (Birch, 1979a, 1979b, 1980). In contrast, the impact that taste has on adults’ food intake is much less apparent. Adults’ taste preferences and aversions are not always direct predictors of food consumption, probably because of weight-related attitudes and concerns with anticipated consequences of food ingestion, such as healthfulness and fat content, and/or costs and ease of obtaining and preparing the food (Birch and Sullivan, 1991).

REPEATED EXPERIENCES AND PHYSIOLOGICAL CONSEQUENCES OF FOOD INGESTION

An influential factor in the consumption of a novel food is how often it has been encountered. The ability to learn to associate the sensory properties of a food with the metabolic consequences of eating that food is widespread in the animal kingdom. Animals use such conditioned preferences and aversions to direct their selection of foods (Forbes, 2001).

According to the “learned safety” hypothesis (Kalat and Rozin, 1973), if the ingestion of a novel food does not have negative post-ingestive consequences it gradually becomes familiar and becomes part of the animal’s diet. Experimental evidence in the rat has shown that energy-dense foods become preferred to low-energy ones after repeated opportunities to ingest those foods and to learn to associate flavor cues with the post-ingestive consequences of eating foods with different energy densities (Booth, Mather and Fuller, 1982; Sclafani, 1990). At least in contexts in which food is scarce, learning food preferences on the basis of energy density is adaptive since it would be advantageous to prefer foods that provide a higher energy amount (Birch, 1998).

On the contrary, if the ingestion of a novel food has noxious consequences, an individual will associate them with its consumption and that food will not be eaten anymore. Food aversion learning (Garcia, Kimeldorf and Koelling, 1955; Garcia and Koelling, 1966), first documented in rats, has been demonstrated in humans (Schafe and Bernstein, 1996) and non-human primates (for a review see Visalberghi, 1994). This robust learning mechanism primarily concerns novel foods, and to a lesser extent familiar ones. In food aversion learning the ingestion of a food that is associated with strong negative experience(s), such as gastrointestinal illness, leads to complete avoidance of the noxious food. The avoidance persists when the food is no longer noxious, and avoidance learning is quicker when the food is novel than when the food is familiar.

For captive capuchins, a food remains unfamiliar (i.e., they respond to it neophobically, see above) for the first few encounters. Visalberghi, Valente and Fragaszy (1998) showed that, after five or six short presentations, several novel foods were eaten to a similar extent as familiar ones. In Visalberghi, Sabbatini, Stammati and Addessi’s (2003) study (see above), in the Post-treatment, after a familiarization period with the novel foods (Treatment), capuchins’ preferences changed from those shown in the Pre-treatment. In fact, capuchins’ Post-treatment choices were significantly correlated with the total energy content rather than to sugar content. It is possible that, after repeated experiences with the 7 foods, capuchins learned to associate the sensory properties of the foods to the consequences of their ingestion and made their choices on the basis of the long-term feedback coming from nutrient content rather than the short-term feedback provided by the foods’ sweetness (see above). The fact that in the long run capuchins’ food choice does not rely only on one source of metabolic energy (sugars), but reflects a wider assessment of the food energy value, may explain the high degree of flexibility of their diets in natural habitats.

Children’s preferences are also shaped by the frequency of consumption of novel foods (Birch and Marlin, 1982; Sullivan and Birch, 1994), whereas just smelling or looking at the food does not have any significant effect (Birch, McPhee, Shoba, Pirok and Steinberg, 1987). Children familiarize not only with the novel food’s characteristics (i.e., its visual appearance, texture, flavor, and odor) but also learn about the negative and positive physiological consequences of ingesting it. In fact, if parents continue to present the initially rejected food and simply require that the child take a very small bite, the food will be accepted eventually (Birch, 1983). Birch and collaborators conducted a series of experiments in which children had repeated opportunities to consume two different versions of the same food (e.g., yogurt) that differed in energy density (for a review, see Birch, 1996). Results showed that three- and four-year-old children learn to prefer foods with high energy density over foods with low energy density by associating different flavors with the consequences of food ingestion (Figure 1). This finding held true both when differences in energy density were produced by manipulating the fat or carbohydrate content of foods; therefore, energy density is the relevant factor for establishing the association, wherever the energy comes from.

Figure 1.

Total amount of drinks with high-caloric- (black bars) and low-caloric-density paired flavors (white bars), ingested during pre- and post-conditioning. Modified from Birch, McPhee, Steinberg and Sullivan (1990).

Figure 1. Comparaison de la consommation totale d’une boisson à forte densité calorique (histogramme noir) et d’une boisson peu calorique (histogramme blanc) avant et après une période de conditionnement aux flaveurs associées à chacune de ces boissons (d’après Birch, McPhee, Steinberg et Sullivan, 1990).

The change in acceptance of novel foods occurs relatively slowly in 2- to 5-year-old children, often requiring as many as 8 or 10 exposures before a significant decrease in neophobia is shown (Sullivan and Birch, 1990). In contrast, for infants (4- to 7-month-old) relatively minimal experience with a food produces a significant increase in food intake. This finding is in agreement with the observation of a lower food neophobia level in infants than in older children (Birch et al., 1998).

DO SOCIAL INFLUENCES SHAPE FEEDING BEHAVIOR?

The acquisition of information about food from group members seems particularly relevant in omnivorous species whose diets often include novel food sources. It has often been assumed that primates learn to identify foods they eat from others and that dietary convergence or diffusion of new feeding habits in wild groups result from social learning (e.g., Kummer, 1971; Nishida, 1987; for a critical review see Visalberghi, 1994). These assumptions stem from the belief that learning from experienced group members is a safer way to select adequate foods—without the risk of getting poisoned or eating foods lacking a good balance of nutrients—than trial-and-error learning.

Since “social dynamics influence the likelihood of social learning” (Coussi-Korbel and Fragaszy 1995, p. 1446), learning about food from group members should especially be expected in tolerant species. Capuchin monkeys exhibit a high degree of inter-individual tolerance, particularly towards infants and juveniles, in the wild (Izawa, 1980; Janson, 1996; Perry and Rose, 1994) as well as in captivity. In captivity, food is sometimes transferred from one individual to another (de Waal, Luttrell and Canfield, 1993; Fragaszy, Feuerstein and Mitra, 1997; Thierry, Wunderlich and Gueth, 1989). Given their relatively high inter-individual tolerance, investigation of social influences on feeding behavior in tufted capuchins is potentially promising (Figure 2).

Figure 2.

Adult capuchin monkeys tolerate youngsters inspecting at close range items that they hold (Photo by Amy T. Galloway).

Figure 2. Les singes capucins adultes montrent une grande tolérance vis-à-vis des jeunes individus qui viennent examiner la nourriture qu’ils manipulent (Photo Amy T. Galloway)

Visalberghi and Fragaszy (1995) showed that capuchins ate significantly more of novel foods when in the presence of group members eating the same novel foods, whereas with familiar foods the increase was not significant. They argued that the social learning mechanism accounting for the increased consumption of novel foods was social facilitation (i.e., the increase in frequency of a behavior already in the species’ repertoire when observing other individuals performing that behavior, see Clayton, 1978). However, only social facilitation of eating what the others are eating can be considered a safe way to learn about a safe diet, and in Visalberghi and Fragaszy’s (1995) study it was not possible to assess whether the occurrence of social facilitation was directed towards the same food the group members were eating, since only one type of food was available to all subjects.

To investigate whether social facilitation of feeding behavior occurs only when the food eaten by group members matches the color of food eaten by a naïve observer, a series of four experiments were carried out in capuchin monkeys (Visalberghi and Addessi, 2000, 2001; Addessi and Visalberghi, 2001; for a review, see Visalberghi and Addessi, 2003, Addessi and Visalberghi, in press). Since primates rely primarily on vision (Jacobs, 1997), in these studies, based on an observer-demonstrator(s) paradigm, the input provided to the observer by the demonstrator(s)’ eating activities was primarily visual. The experimental setting included a transparent panel separating the naïve observer (who was in one cage) from its demonstrator(s) (who was/were in the adjacent cage). The transparent panel allowed the observer to see its demonstrator(s) (and vice versa) but prevented food transfers, as well as the possibility for the observer to smell demonstrator(s)’ muzzles and/or food.

In Experiments 1 and 2, the observer received a novel food in a box attached to the panel, and the demonstrators received a food in a larger box on their side of the panel. The foods were mashed and colored with strikingly different colors, or with the same color, according to the experimental condition. Results showed that the observer’s consumption of novel foods was significantly lower when the observer was alone than when the observer could see its demonstrators eating a different colored food (Experiment 1). In addition, there was a significant correlation between the number of demonstrators who were eating and the number of observers eating at the same time (Visalberghi and Addessi, 2000; Addessi and Visalberghi, 2001). In Experiment 2, the observers ingested to similar extents both the food that matched the color of the one eaten by the demonstrators and the food that did not match the color of the one eaten by the demonstrators (Visalberghi and Addessi, 2001). Therefore, because the demonstrators’ eating behavior increased the observer’s acceptance of novel foods regardless of whether the demonstrators’ and observers’ foods were the same, the hypothesis that the social context fosters learning about a safe diet seems unsupported.

However, since in general animals are rarely faced with only one food source at the time, Experiments 3 and 4 were carried out to assess whether, when given a choice between two novel foods, only one of which matches in color the food that demonstrator(s) are eating, a capuchin would match the demonstrator(s)’ choice. In Experiment 3, the observer was presented with a choice between two versions of a novel food, one of which matched the color of the food eaten by demonstrators, and one that did not. Results showed that observers did not preferentially choose the matching-color food rather than the non-matching one: observers’ latencies to eat and amount of food ingested did not differ between the matching- and the non-matching food (Visalberghi and Addessi, 2001). In Experiment 4, a demonstrator was presented with two foods of the same colors as the novel foods presented to the observer, but the demonstrator ate only one food and systematically avoided the other one. Regardless of this apparently more salient input provided by the demonstrator to the observer, once again observers’ choices were not affected by demonstrators’ behavior (i.e., observers ate similar amounts of the two foods) (Addessi and Visalberghi, in press) (Figure 3).

Figure 3.

Experimental paradigms used in a series of experiments carried out to investigate social influences on the acceptance of novel foods in capuchin monkeys (see text).

Similarly, in Visalberghi, Sabbatini, Stammati and Addessi’s (2003) study, although capuchins significantly changed their initial food preferences after repeated experiences with the novel foods during the Treatment, the Social or Individual condition in which individuals encountered the foods during Treatment did not affect their change in food preferences. On the contrary, during the Post-treatment, subjects’ preferences were more alike than during the Pre-treatment.

In conclusion, even though capuchins should have been a promising species in which to look for social influences on feeding behavior, social influences on feeding seem generic and not goal-oriented and specific. The above experimental findings do not support the common assumption that monkeys learn from one another to identify what to eat.

For young children, eating is typically a social occasion (Birch, 1983) and the social context in which food is presented strongly influences the child’s food acceptance patterns. When foods are given to children in positive social contexts (as rewards, or paired with positive social interaction with an adult), children’s preferences for those same foods are enhanced (Birch, Zimmerman and Hind, 1980).

Social facilitation of feeding behavior has been repeatedly demonstrated in young children. Harper and Sanders (1975) conducted a study to assess whether (1) adults’ eating unfamiliar foods would induce young children to accept similar items; (2) toddlers are more susceptible to social influences than older children. Eighty children participated in the study: forty toddlers and forty 3½-yr-olds. Half the subjects in each age group were offered a novel food (Offer-Only condition); the other half was offered a novel food after an adult already had begun to eat (Adult-Eat condition). Within each subgroup, half the subjects received the novel food from the mother, and half from a friendly unfamiliar adult. Results showed that there was a consistent tendency for young children to sample an unfamiliar food more readily when an adult was eating it than when the food was merely offered. Both mothers and unfamiliar adults were successful models, even though mothers’ influences were more effective than unfamiliar adults’. The older children appeared to respond in the same way as the toddlers. However, similarly to Visalberghi and Fragaszy’s (1995) study, it was not possible to assess whether social facilitation was specifically directed towards the same food the adults were eating, since only one type of food was available to adults and children.

In another study, Birch (1980) investigated the influence of peers’ modeling of food selections and eating behaviors on preschoolers’ food preferences and consumption when more than one food was available at the same time. Thirty-nine children’s preferences for vegetables were assessed. Target children (N = 17), who preferred vegetable A to B, were seated with 3 or 4 peers with opposite preference patterns. Children were served preferred and non-preferred vegetable pairs at lunch and asked to choose the preferred one. On day 1, the target child chose first, while on days 2, 3 and 4 peers chose first. Results showed that target children shifted from choosing their preferred food on day 1 to choosing their non-preferred food by day 4 (McNemar chi square = 5.82, p< 0.05; see Figure 4). Moreover, 12 of 17 target children increased the consumption of non-preferred vegetables, whereas only less than half of the peers did so. Therefore, exposing children to peer models that were selecting and eating the target child’s non-preferred food was sufficient to increase the number of choices for the initially non-preferred food by the target children, even in the presence of an initially highly preferred food. Younger children (three-year-olds) were more affected by peer modeling than older children (four-year-olds).

Figure 4.

Percentage of choices directed towards the non-preferred foods by target children (black bars) and peer models (white bars), during each of the 4 days of the experiment. Modified from Birch (1980).

Figure 4. Pourcentages des choix des aliments non préférés par les enfants observateurs (histogramme noir) et par leurs pairs constituant les modèles (histogrammes blancs) au cours des quatre jours de la période d’expérimentation (d’après Birch, 1980). Le premier jour, les observateurs choisissent en premier, sans référence à un ‘modèle’; puis, ayant observé les démonstrateurs qui choisissent en premier les jours suivants, le comportement des enfants observateurs varie significativement.

The success of the peer-influence procedures reported above suggests that if children were routinely exposed to other children with preferences different from their own, the set of foods they would learn to accept would be enlarged. On the contrary, encouraging intake of healthy foods, such as fruits and vegetables or forcing a child to eat a food to obtain a reward are ineffective strategies (Birch, Marlin and Rotter, 1984). In fact, parental control impedes the acquisition of children’s self-control of food intake, and may foster the development of unhealthy diets and overweight in childhood (Birch, 1998).

CONCLUSIONS

Omnivorous species have to learn continuously about what is safe to eat. Experimental evidence showed that in capuchins (1) individual experiences based on the feedback obtained from different foods guide the establishment of preferences towards novel foods, (2) the short-term sensory feedback provided by taste perception precedes the long-term feedback coming from total energy content in influencing food preferences, and (3) repeated individual experiences are sufficient to determine food preferences similar to those individuals may acquire when consumed together with group members eating the same foods.

Moreover, even though capuchin monkeys are strongly interested in the novel food another monkey is holding or eating (Visalberghi and Fragaszy, 1995; Drapier, Addessi and Visalberghi, 2003), naïve capuchin monkeys do not seem to learn about food palatability through interactions with more experienced group members. However, since social facilitation of eating affects the acceptance of novel foods, it can be considered a quicker way, compared to individual learning, to overcome food neophobia. In fact, feeding together with group members might increase the chances that a naïve individual’s feeding activities are channelled toward those of its group members.

Similarly, during the first 5 years of life, human children have to learn an enormous amount of information about food. Most of this information is acquired in absence of explicit teaching (Birch and Fisher, 1995), by means of processes of associative learning. Associative learning is particularly important for learning of evaluative and emotional responses, such as food preferences (Martin and Levey, 1978). In this last case, associations may be established between foods and the positive or negative cues present in the social contexts in which eating occurs, or may arise as a consequence of eating (Birch, 1998).

Therefore, children’s preferences for energy-dense foods, and especially for high-fat foods, might be due to learned preferences, acquired via associative learning as a result of repeated opportunities to consume these foods and to experience the effects of their ingestion. As a consequence, an environment that provides many opportunities to experience high-fat, energy-dense foods and few opportunities to learn to like fruits and vegetables could be described as obesigenic (Birch, 1998).

In conclusion, given the above information we think that the research findings and the methods used to investigate the factors affecting the acquisition of the diet in capuchin monkeys and children can inform both programs of research. On the one hand, the many similarities between the two species can demonstrate how experience with food is critical for shaping food preferences. On the other hand, the major difference between the two species, i.e., the fact that social influences are so much more powerful and specific in children than in capuchins, can provide insight about the reasons why in our society food is becoming such a source of concern. Unfortunately much of this worry about food is not because people are starving, but because they respond to food by over-eating or under-eating. The research with children has benefited the capuchin research by focusing attention on critical questions that are ripe for study. Because capuchins monkeys are not known to experience problems related to over- or under-eating, (despite the fact that they are a social species), they might be able to provide useful clues concerning human ingestive behavior.

RESUME

L’acquisition des comportements alimentaires permettant de couvrir les besoins nutritionnels et d’éviter les produits toxiques est essentielle pour la survie de tout organisme vivant. A un stade précoce, chez le jeune humain, de même que chez les primates non humains, les choix alimentaires ne sont pas influencés par le contexte culturel qui prévaut souvent lorsque des adultes déterminent leurs choix en fonction d’attitudes et de croyances relatives à une image corporelle valorisante. De ce point de vue, le comportement alimentaire du singe capucin et celui de l’enfant humain présentent des caractéristiques communes, malgré les 40 millions d’années qui séparent l’évolution de l’homme de celle des singes platyrrhiniens. Une comparaison de ces comportements précoces peut nous apporter des éléments de réflexion. En effet, de même que l’espèce humaine, le singe capucin est omnivore, longévive avec une longue phase de développement, et son encéphale particulièrement développé par rapport à la taille corporelle est, à la naissance, immature du point de vue neuronal. L’objectif de la présente étude est d’abord de déterminer les différents facteurs qui peuvent influencer les choix alimentaires et le comportement d’ingestion chez les espèces omnivores. En second lieu, la discussion des résultats, situés dans le contexte d’autres études sur les singes capucins et sur les enfants humains, permet d’élargir, dans la mesure du possible, la perspective comparative.

Le succès d’une espèce omnivore dépend de sa propension à explorer et à tester des aliments nouveaux afin de les inclure dans son régime, mais également de sa prudence vis-à-vis de ceux-ci pour éviter d’ingérer des substances toxiques. Une solution partielle à ce “ dilemme des omnivores ” est la néophobie alimentaire, c’est-à-dire la réticence à consommer des aliments nouveaux. Aussi bien en captivité que dans leur milieu naturel, les singes capucins présentent ce comportement néophobique. De même, les enfants humains ont tendance à se méfier des aliments nouveaux, notamment pendant la période de sevrage où ils passent du régime lacté à une alimentation diversifiée. Toutefois ce comportement prudent vis-à-vis d’aliments qui pourraient s’avérer toxiques doit faire l’objet d’un compromis avec la nécessité d’une alimentation variée.

Nous savons par ailleurs que la perception gustative qui permet aux primates d’éviter d’avaler des substances potentiellement toxiques est également une réponse bien adaptée aux choix des aliment bénéfiques et caloriques. Au cours des premières phases de la vie, les goûts plus ou moins appréciés constituent les principaux déterminants de la prise alimentaire, et les jeunes enfants acceptent sans problème les aliments sucrés et refusent les produits amers ou acides. Habituellement, ils tendent à consommer de grandes quantités des aliments préférés, en l’absence du contrôle exercé par les adultes.

Bien que les réponses aux stimuli gustatifs du capucin nouveau-né n’aient pas été décrites, les réflexes gusto-faciaux et les réponses ingestives à des solutions sucrées sont vraisemblablement équivalents à ceux qui ont été mis en évidence chez d’autres espèces de primates. Chez les jeunes capucins et chez les adultes, la perception gustative, qui déclenche une sensation immédiate agréable ou désagréable, détermine les préférences alimentaires. Le goût des sucres solubles n’étant généralement pas associé à des effets toxiques, chez une espèce omnivore comme le capucin, l’intensité de la saveur sucrée apparaît comme un critère bien adapté pour le choix de nouveaux aliments afin de diversifier le régime sans prendre de gros risques.

Parmi les facteurs qui déterminent la consommation d’un nouvel aliment, la fréquence de l’exposition à cet aliment est également cruciale. Pour un capucin, en situation expérimentale, les premiers contacts avec un aliment nouveau entraînent des réponses néophobiques. Après une exposition répétée, le sujet associe les propriétés organoleptiques de l’aliment nouveau avec les conséquences de son ingestion. Le choix s’opère alors en fonction d’une perception à long terme qui résulte du contenu énergétique de l’aliment plutôt que de la perception gustative immédiate en relation avec la teneur en sucres solubles. De la même façon, chez les enfants humains, les préférences alimentaires sont acquises, après des expositions répétées leur donnant l’occasion de goûter un nouvel aliment, par les processus d’association qui mettent en jeu la perception des effets post-ingestifs.

Chez les espèces omnivores, les contacts entre les membres d’un groupe facilitent l’acquisition de connaissances sur les nouveaux aliments. Un singe capucin consommera une quantité d’autant plus grande d’un aliment nouveau qu’il se trouve en présence d’autres individus qui consomment cet aliment. Cependant, la consommation d’un aliment nouveau peut être accrue par le simple fait d’observer un congénère en train de manger, même si le “démonstrateur” consomme un aliment différent. Chez le capucin, la facilitation sociale du comportement alimentaire n’est donc pas spécifique d’un aliment déterminé. Elle peut être considérée comme un moyen plus rapide que l’apprentissage individuel pour surmonter la néophobie.

Chez les jeunes enfants humains, le repas se situe toujours dans un contexte social qui influence fortement l’acceptation des aliments et les préférences alimentaires. La facilitation sociale est bien démontrée car un jeune enfant essaye un aliment non familier plus rapidement lorsqu’un adulte le consomme que lorsque cet aliment lui est simplement offert. Les mères, ainsi que des adultes non familiers, constituent des modèles efficaces, même si ces derniers n’ont pas autant d’influence que la mère. De la même façon, les enfants en contact avec leurs pairs, considérés comme modèles et consommant des aliments nouveaux, sont amenés à consommer et à préférer ces aliments.

En conclusion, les deux approches concernant l’acquisition du régime, d’une part chez les jeunes singes capucins, d’autre part chez les enfants humains, s’enrichissent mutuellement par les résultats qu’apportent des méthodes complémentaires. Les similitudes observées entre les espèces montrent le rôle crucial de l’expérience dans l’établissement des préférences alimentaires. Cependant leur différence fondamentale à propos du contexte socioculturel qui, chez les enfants humains, détermine des choix de manière beaucoup plus spécifique que chez les capucins, pourrait nous éclairer sur les bases sociales à partir desquelles l’alimentation est perçue dans nos sociétés.