Discussion

As demonstrated by the above examples of nonhuman primate communication, one could postulate that there are certain aspects of human vocal production which are similar. Additionally, one can hypothesize that the ability to vary the arrangement of vocal elements, and to produce meaning differences, as found in gibbon duets, chimpanzee long calls and vervet alarm calls, may be the source of language origins. However, it is important to remember that nonhuman primates are not ‘unevolved’ human beings. Humans and nonhuman primates are completely different species. Behavioural and physiological traits from one may be used as an analogy for the other species. However, it makes no sense to state that chimpanzees are more evolved than macaques, nor that the vocal capabilities of humans are ‘more complex’ than that of other nonhuman primates.

Instead, one can examine which fundamental traits define the communication of nonhuman primates, and other animals: 1) the ability to recognize the vocalizations of its particular species, as distinct from other environmental noises; 2) the ability to recognize the vocalizations of kin; 3) the ability to recognize the vocalizations of allies and ‘strangers’; 3) the capacity to learn new vocalizations; 4) the capacity to discern ‘referential meaning’ contained within a species specific vocalization (i.e. the difference between a mating call and an eagle alarm call); and 5) the capacity to alter vocalizations without altering the initial referential meaning.

It is important to remember that human beings are not born as adults. We have a long, steep learning curve which takes years to master. Secondly, vocal communication is only one form of communication. Both humans and nonhuman animals rely upon visual and gestural communication, as much, if not more, than upon vocal communication. Additionally, the majority of human language is little more than ‘grunts’ of acknowledgment or other noises (umm) which fill conversational silences. Although Bickerton (1990) would say that these ‘words’ fulfill a grammatical function by occupying a ‘position’ in a sentence, one could just as simply state that it is no more complex than a chimpanzee long call response. Finally, defining human language as ‘complex’, since it is not bounded in space or time, is a bit of a stretch. Languages are fluid, their complexity arises once their vocabularies increase, which has more to do with the consensus regarding novel acoustic sounds than the so called grammatical rules of a language. In other words, if the constraints of the hypotheses fail when applied to humans, why are they employed on nonhuman primates?

Vocal Responsiveness

In human conversation, individuals appear to take turns in language production, which is characterized by acoustically distinct sounds (Arcadi 2000:206). In contrast, although simple vocal exchanges involving distinct calls have been documented in some primate species, these calls appear to be acoustically similar (ibid.). Mitani and Brandt (1994:250) concluded that there is no evidence that chimpanzee vocal behaviour is different from that of any other primate. However, others (i.e. Burling 1993; Ujhelyi 1998) have stated that chimpanzee vocal behaviour is more ‘sophisticated’ than that of other primates, and therefore may provide insights into the origins of language. However, acoustic analysis has revealed that some primate calls which sound the same to human observers are in fact distinct vocalizations that are employed in different contexts and elicit different behavioural responses (Arcadi 2000: 218; Zuberbühler et. al. 1997:601; Cheney and Seyfarth 1982:748).

Seyfarth and Cheney (1997) found that the ‘grunts’ of female baboons served as a reconciliatory signal because they reduced the anxiety of lower ranking females’ after aggression. In vervet monkeys, Cheney and Seyfarth (1982) have identified four different social situational ‘grunts’: upon encountering a dominant conspecific, upon encountering a subordinate conspecific, to a conspecific moving to an open area and to vervets who are not members of the group. Seyfarth and Cheney (1997) also observed that vervet ‘wrr’ calls were used to indicate that a neighbouring group had been seen and ‘chutter’ calls indicated that an intergroup encounter had become aggressive.

In 1996, Arcadi spent 53 days observing a small group of chimpanzees (10 adult males, 8 central adult females and 9 peripheral adult females) which had never been provisioned, in Kibale National Park, Uganda. Arcadi (2000:213) found that wild chimpanzees vocalize at low rates, tend not to respond to calls that they hear, and when they do respond, they tend to give calls that are similar to the ones they have just heard. She noted that calling rates were higher when other calls were audible, and temporally clumped calling within and between subgroups typically involved either chorusing or counter calling with calls of the same type as those just heard (Arcadi 2000:217).

Social Learning

Learned vocalizations function as an indicator of group membership. Chimpanzees and modern humans live in social groups characterized by within group cooperation and competition between groups. Such social systems put a premium on reliable indicators of group membership, vocal or otherwise (Tomasello and Call 1997). An alternate possibility is that vocal learning is just one example of a domain general mimetic ability in modern humans (Fitch 2000).

Janik and Slater (2000:8) define three different types of social learning: contextual, production and vocal. Contextual learning affects the behavioural context, both usage and comprehension, of a signal. Production learning refers to instances where the signals themselves are modified in form as a result of experience with those of other individuals. Vocal learning is defined as production learning in the vocal domain. It can affect one or more three systems which involve different levels of control over sound production (ibid.): respiratory, phonatory and filter. According to Janik and Slater (2000), contextual learning and respiratory production both preceded the evolution of phonatory and filter production learning.

At the same time, the most important primate specific tool for expressing emotions is the facial gesture. Producing modifiable facial gestures has an immediate communicative function” (Ujhelyi 1998:180). As secondary result, acoustic variations, arise if an animal changes its facial gesture during vocalization. Primates have mobile, nonattached upper lips, which enable them to produce different facial expressions. Different facial expressions include lip configurations which form barriers to the passage of air (Burling 1993:30). Thus they may result in different acoustic outcomes. Although a human like vocal tract is absent, the face will be the main tool for producing articulatory variants (ibid.).

Since most vertebrates can distinguish the vocalizations of different individuals (parent/offspring, conspecifics, strangers), formants, the ‘shaping’ of sound due to physiological constraints, plays an important role in individual identification (Firth 2000:263; Burling 1993). Nonhuman primates also have formants in their calls, which vary with context (Janik and Slater 2000:9). Formants might also provide an indication of the body size of the vocalizer. Firth (2000:263) has proposed that there is correlation between vocal tract length and body size in humans and monkeys. However, formant cues are different than that of vocal pitch, which is not correlated with body size in humans (Lieberman and McCarthy 1999:488).

One could presume that early sound making would have been accompanied by rudimentary vocal facial displays such as those produced by modern nonhuman primates. Later changes, building upon inherent preadaptations, expanded from the strictly prosodic domain into the articulatory domain, and subsequently, the discovery that these sound making movements could be combined in various ways to produce a range of distinctive phonetic forms (words).

Duets

A duet is defined as long calls or songs in which both sexes of a monogamous pair produce loud sounds in an interactive manner, performing a mutually cooperative and coordinated display (Arcadi 2000; Ujhelyi 1998). In some primate species, mated pairs sing ‘duets’ and neighbours ‘counter sing, in an alternating but apparently timed manner. For example, chimpanzee long calls are thought to maintain connections with group members, gibbon duet songs are performed at given times of the day, and indris duets are produced only during the breeding season (ibid.). In gibbon duets, the contribution of males and females to song display show a rather rigid and uniform pattern (Arcadi 2000). Although there are some instances in which song transfer may occur: for example, when a female becomes widowed, she may adopt and perform the male song and so produce a pseudo duet the strong sexual differences in song structure are likely to be genetically programmed (Ujhelyi 1998).

Despite the strong correlation between the duet performance and monogamy, coordinated call display does exist in great apes. The common chimpanzee males often call together, while in bonobos, a male-female pair duet occurs (Ujhelyi 1996, 1998). In both chimpanzee species, duetting or chorusing can be heard all day in relation to different activities (Ujhelyi 1998:184). It has been proposed that duetting is a definitive feature of stable monogamous and territorial primate species, and that the function of duetting may be the maintenance and reinforcement of the pair bond (Ujhelyi 1998:185). Differences in partner preferences are due to differences in group structures between the two chimpanzee species. Ujhelyi (1998) observed that bonobos show a high degree of synchronization between vocalization of different individuals. Ujhelyi (1996, 1998) further postulates that the capacity of duet performance might be a remnant of the earlier monogamous stage, and altered to fit in the current way of life.

Mitani and Brandt (1994:250) observed that chimpanzee males attempt to match the acoustic characteristics of each other’s vocalizations when calling together. Single males appear to alter the acoustic structure of their calls when chorusing with different partners. This tendency results in large variability in call types on the one hand, but homogenization in call repertoire of the group on the other hand (ibid.). According to Marshall et. al. (1999:826), the call repertoire being acquired by a single male may contain a large number of variants mostly acquired via social learning, while the call repertoire itself is not exclusive to a specific individual.

Long calls

Territorial song marks the territory of a group and serves to maintain spacing between members of neighbouring groups. Due to its acoustic nature, it is impossible to mark territory directly, instead the presence, identity and location of the territory owner are broadcast (Ujhelyi 1998:184). Ujhelyi (1998:185) considers this to be a representational function. According to Ujhelyi (1998:180) this territorial behaviour establishes lexical syntactic capacity. If the labeling channel is an acoustic one, and the primary sounds are limited and genetically fixed, then differences in signs “can only be achieved by compositions of the invariant elementary sounds and by varying their arrangement.” In contrast, species capable of producing within call acoustic variants live in large groups with complex social interactions (ibid.).

All the primate species, excluding African apes, which long, variable calls share a common feature in their social behaviour, namely monogamous territoriality (Arcadi 2000:216; Ujhelyi 1996:74). In contrast to other territorial mammals that rely upon olfactory marking, most nonhuman primate species mark and defend individual territory by acoustic signs (ibid.). These distinctive loud calls are given by males as territorial displays, and these can elicit similar calls from one or more conspecific (Ujhelyi 1996, 1998; Tomasello and Call 1997). These long calls or “songs” are displayed without any overt external stimulus and have some musicality in nature. It has been proposed that the songs of different species represent different degrees of complexity (Burling 1993).

Long calls are built from smaller, stable, clearly distinguishable units, and exhibit individual variation over time (Burling 1993; Ujhelyi 1996, 1998). The units of these long calls are ‘traditional’ communicative signals (i.e. alarm or contact call) which are combined in different arrangements in different compound calls. The number of elementary units in long calls differs across species, and acoustically different songs can be created by changing the number, type and position of elements (Ujhelyi 1998:179). According to Mitani and Marler (1989:43) the gibbon song may be divided into distinct vocal elements. Based upon seven variables (duration, maximum frequency, minimum frequency, frequency range, start frequency, end frequency and number of frequency inflections), 13 basic note types can be distinguished (ibid.) The songs are then built up from these notes. The songs can be varied using different type, number and positions of elements, segments or notes. Hence, the song repertoire of an individual male may be rather large.

Both the common chimpanzees and bonobos have long calls, which can be divided into some acoustically distinct segments, similar to gibbon songs (Mitani and Marler 1989; Clark and Wrangham 1994). Chimpanzee vocalizations are highly graded with many variants used in a wide range of contexts (Arcadi 2000:205). It has been noted that these vocal sequences can be long and involve many call types (ibid.; Ujhelyi 1998). Additionally, extended vocal exchanges between individuals out of visual contact are common (Arcadi 2000:206). Chimpanzees and bonobos also emit long, compound calls (pant hoot and high hoot, respectively) which can be divided into acoustically distinct segments (ibid.). Although chimpanzees do not change the order of the four fundamental units of the long call, they insert individually selected vocal elements into different positions of the call (Burling 1993). Chimpanzee males often give the long call together, during which they attempt to match the acoustic characteristics of each other’s vocalizations. The matching tendency shows that call variants can be learned (Arcadi 2000:206; Ujhelyi 1998:185-186).

Ujhelyi (1998:179) has proposed that this type of call variant production “may represent phonological syntax since the altered parts of the call do not possess their own meaning independent of the call”. In other words, this type of call production may represent an intermediate stage between animal communication and language. As a result of living in a large social group, the ability to create syntactically different calls was enhanced. A call repertoire emerged which contained a large number of call variants at group level available for each group member via social learning. Ujhelyi (1996, 1998) believes that this type of animal call is different from ordinary animal communication since it apparently demonstrates some features of human language.

Zuberbühler et. al. (1997:601) found that the long distance calls of diana monkeys function in perception advertisement as well as within group semantic signals that denote different types of predators. Subjects were a group of 20-25 individuals (1 male, 5-7 adult females, subadults and infants) in the Taï National Park, Côte d’Ivoire. It was observed that diana monkeys show age/sex dimorphism in the vocal repertoire (Zuberbühler et. al. 1997:591). Adult females, subadults and juveniles accounted for most of the vocal activity in the group, and are responsible for the following vocalizations (ibid.): contact call, trill, alert calls (leopard, eagle) and agonistic calls directed in both intra and intergroup interactions. Male diana monkeys tended to restrict their vocal communication to long distance calling to which females responded with they own, acoustically different, alarm calls (ibid.). According to Zuberbühler et. al. (1997:601), the long distance calls in nonhuman primates show acoustic specialization. Calls are structurally stereotyped and are given repeatedly (ibid.).

It seems that it is just this territorial behaviour which first established the linguistic capacity. If the labeling channel is an acoustic one, and the primary sounds are genetically fixed. Then only by varying the elementary sounds can sign differences be achieved. Consequently, those individuals who are capable of linking, repeating, and combining these elements get selective advantages. It can be shown that some of the notes of gibbon song occur independently of song, in another context reaction to encounters (Mitani and Marler 1989). These simple elements function in ordinary communicative situations. The combination of available elements resulted in a variable set of songs which became suitable for territorial marking.

According to Zuberbühler et. al. (1997:601), in rain forest habitats, where visibility is generally poor, the acoustic domain may provide the most efficient means by which a prey animal can communicate to a predator. In the case of diana monkeys, calls are given only to hunters which surprise their prey (leopards, eagles) and not to hunters that pursue their prey (chimpanzees, humans) (Zuberbühler et. al. 1997:602). Secondly, calls given in the this contexts are regularly combined with approaching the predators both under experimental and natural conditions (ibid.).

Alarm Calls

Many nonhuman primate species employ various types of ‘referential’ calls to conspecifics. These calls can be classified as food calls, predator alarm calls, and calls for aid (recruitment calls) (Tomasello and Call 1997). Gouzoules et. al. (1984:182) observed that juvenile rhesus macaques, while being attacked by another individual, used one of five different calls to recruit support from relatives. Gouzoules et. al. (1984:183) further grouped these calls as ‘noisy screams’, which are employed by high ranking individuals when there is physical contact, and ‘pulsed screams’ which are used when the attacker is a relative. When recordings of these calls were played during experimental testing, it appeared that individuals were responding to the acoustic properties of the calls, and not the behaviour or emotional arousal of the caller (Gouzoules et. al. 1984:190).

Gouzoules et. al. (1984:190) proposed that the ‘responder’ employed these acoustic properties to determine both the identity of the aggressor and what type of aggression was occurring. Gouzoules and Gouzoules (1995) later repeated this playback experiment with pigtail macaques, since they also appear to employ recruitment screams during agonistic contexts. They observed that the acoustic properties of these calls and the aspects of the agonistic situation differed from that of rhesus macaques (Gouzoules and Gouzoules 1995:449). However, they believed that there is no evidence that the calls of pigtail macaques encode the kinship status of the aggressor (ibid.).

The alarm calls of vervet and diana monkeys have been thought to ‘refer’ to features of the environment (Cheney and Seyfarth 1980, 1990, 1990a). Cheney and Seyfarth (ibid.) state that vervet monkeys employ three distinct predator specific alarm calls (leopard, eagle, snake). A loud barking call is given for leopards; a short, ‘cough like’ call is given for eagles; and a ‘chutter’ call is given for snakes (Cheney and Seyfarth 1990; Hauser 1996). Each call (stimulus) elicits a different escape (behavioural) response on the part of the receiving conspecifics (leopard alarm calls - run up the nearest tree; eagle alarm calls - look up, run into the bushes; snake alarm - stand on hind legs and look down at the ground) (ibid.).

However, it is assumed that this type of referential acoustic behaviour differs from the human usage of words for the following reasons (Seyfarth and Cheney 1997:252): 1) alarm call meaning appears to be limited to the connection between referent and sound; 2) alarm calls are difficult to specify; and 3) may be the result of simple conditioning. Seyfarth and Cheney (1997:252) state that although vervet calls function in a rudimentary semantic manner, it is uncertain whether vervets recognize the referential relation that exists between their calls and features of the environment. Additionally, it is uncertain whether or not this vocalization is interpreted as a representation of the caller’s knowledge of conspecifics (ibid.; Hauser 1996:413).

The alarm calls of male diana monkeys show consistent differences in acoustic and temporal structure depending on whether they are given to leopards or eagles (Zuberbühler et. al. 1997:602). According to Zuberbühler et. al. (1997:602), the most salient feature of these alarm calls, the number of syllables, did not seem to be sufficient for an unambiguous identification (for the human observers). The alarm call for leopards seemed to have fewer syllables than the alarm call for eagles, although there appeared to be some overlap (ibid.). The female and juvenile diana monkeys responded in qualitatively and quantitatively similar ways to both the male’s call to a predator and to the predator that typically caused that call, in both playback experiments and natural conditions (ibid.). Zuberbühler et. al. (1997:602) concluded from this observation that that these calls contain semantic information. Like vervets and diana monkeys, playback experiments with ring tail and ruffed lemurs indicate that they employ referential calls in specific, aerial or terrestrial, predator situations (Macedonia 1990).

In the case of the great apes, there have been no systematic studies with regard to how individuals understand or employ alarm or recruitment calls (Tomasello and Call 1997). Almost all of the systematic research concerns the food calls of chimpanzees: pant hoot, food grunt and ‘food-aaa’; but only the first two have been systematically studied (ibid.). Clark and Wrangham (1994:199) found that pant hoots given by food discoverers did not increase the frequency with which conspecifics arrived at the food site. Hauser et. al. (1993:818) found that chimpanzees used pant hoots in addition to food grunts when the amount of food was large. However, the function of pant hoots is uncertain, since it has been observed in other social situations, such as excitement or the announcement of an individual’s location (Mitani and Nishida 1993; Tomasello and Call 1997). According to Tomasello and Call (1997:257), pant hoots do not indicate the discovery of food since captive chimpanzees pant hoot in full view of conspecifics, even though all individuals are aware of the food source.

According to Hauser (1996) and Tomasello and Call (1997), the type of referential behaviour described in the preceding paragraphs, may be due to an innate mechanism, and does not necessitate any understanding of ones conspecifics. This mechanism enables these particular animals to generate specific noises in the presence of particular visual or emotional stimuli and to respond in certain ways to particular acoustic or visual stimuli (ibid.). However, this would be true of any type of behaviour, once it is broken down into components and taken out of context. One of the problems with this type of reasoning is that it ignores the overall picture (referential communication) in favor of a behavioural model.

Girneys

Most of the research conducted on free ranging nonhuman primates has addressed alarm calls. However, Locke (1998:194) has suggested that a more interesting class of vocalizations would be produced by contented animals, vocalizing quietly among themselves in a family or small group situation. One type of vocalization that is not screamed, but uttered by various species of monkeys is the girney. Girneys are most frequently produced by mothers who are interacting with other mothers or juvenile females. Girneys seem to be “produced behind closed lips and resemble the sound of an individual who is talking with food in his mouth” (ibid.). Girneys may be issued interchangeably with lip smacking, and are common among animals which live in small, intimate groups. From a physical standpoint, they lack the “rhythm of lip or tongue smacking, but are typically phonated” (Locke 1998:195). Locke (1998) has proposed that if the phonatory aspect of girneys were combined with the pulsatile character of lip and tongue smacking, a human like type of sound making could have been achieved.

Part IV: Nonhuman Primate Communication

It is thought that animal communication and human language(s) have fundamental differences in their structures and functions (Fitch 2000; Janik and Slater 2000; MacNeilage 1998; Beaken 1996:103; Hauser 1996; Ujhelyi 1996). According to Tomasello and Call (1997:232) animal communication is a “ritualized social act designed to induce others to act via their own self directed powers”. These signals are thought to express the animal’s emotional states, which in turn motivates the resultant behavioural actions of conspecifics in given circumstances (Hauser 1996; Tomasello and Call 1997). Furthermore, in animal communication, a rather limited set of messages appears to be transmitted, which are in general, genetically fixed (ibid.). However, there seem to be several exceptions, for example, the existence of different vocal signs for different predators in vervet monkeys (Cheney and Seyfarth 1990, 1997; Hauser 1996; Tomasello and Call 1997).

Although some researchers believe that symbolism is found among primates, it is not thought to be necessarily found in their natural communication system (Aitchison 1998:20). It has been proposed that the alarm calls of vervet and squirrel monkeys may represent an intermediate stage en route to symbolization (Tomasello and Call 1997; Hauser 1996). Human raised chimpanzees, who have been taught language like systems, can use signs as symbols, yet do so mainly when they require something (ibid.). However, the realization of the potential power of names, and the subsequent desire to label everything, has not been found in these experiments, although it occurs in human children between the ages of one and two (Aitchison 1998:21; Tomasello and Call 1997).

Human speech has also been distinguished from nonhuman primate gesture call systems by virtue of its representational level. According to Knight (1998:69), linguistic reference is not a direct mapping from linguistic terms either to perceptible things or to intentional states. Instead the mapping is from linguistic terms to communal constructs and representations established in a structured discourse (ibid.). Human vocal activity may have evolved from instinctive animal vocalizations (Beaken 1996:103). Specifically those of social interaction and emotional expression (akin to those of infants). However, the early vocalizations of infants are quite different from the speech patterns of adult human beings.

Ulbaek (1998:33) believes that language evolved from animal cognition, not communication. Ulbaek (1998:34-46) claims that the following are examples of cognitive abilities in nonhuman animals specifically, primates (Ulbaek 1998:34-36). : 1) tool making and usage (termite ‘fishing sticks’, ‘hammer and anvil’ stones for cracking nuts); 2) cognitive maps (knowledge of their territory to plan routes to food areas); 3) learning through imitation; 4) social knowledge (dominance hierarchies); 5) deception; 6) theory of mind (knowledge of the intentions of others); and 7) ability to learn language like systems (based on experimentation).

The main argument against nonhuman primate communication as a precursor of modern human language is the lack of conclusive evidence that nonhuman primates possess a theory of mind. The theory of mind hypothesis states that humans attribute mental states, such as knowledge and beliefs, to others, and that there is a recognition of the causal relationship between mental states and behaviour (Dunbar 1998; Worden 1998, Tomasello and Call 1997; Seyfarth and Cheney 1997:250; Gibson and Ingold 1993). Nonhuman primates appear to recognize one another as individuals, know all about each others’ kin and alliance relations, and can rapidly learn simple rules about who will do what in specific circumstances (Worden 1998:151; Tomasello and Call 1997, 1993). To have social intelligence, primates need to do three things (ibid.): 1) represent in their minds information about social situations, past and present; 2) to learn and represent internally, the causal regularities whereby one social situation leads to another; and 3) to combine knowledge of the present social situation with knowledge of causal regularities to predict what may happen next.

Based upon the criteria for theory of mind, nonhuman primates may have communication systems which are semantic, but do not qualify as a ‘language’ since they are not ‘intentional’ (Tomasello and Call 1997; Hauser 1996). While it appears that most primates do not possess a theory of mind, the picture in great apes is unclear. Some evidence from field studies suggests that they do, while laboratory studies are more negative (ibid.). However, since these laboratory studies are modeled upon similar experiments with preschool human children, and require a verbal response to correlate observed behaviours, the lack of evidence for theory of mind in nonhuman primates is questionable (Aitchison 1998; Tomasello and Call 1997). Since it is known that it takes several years before human children ‘acquire’ a theory of mind, and there is, as yet, no way to know with any certainty what nonhuman animals ‘intend’ or ‘mean’ when they vocalize, this section will focus upon the interpretations of observed and playback experiments of nonhuman primate vocalizations.

It has been suggested that there is no living species which demonstrates an intermediate stage of language evolution (Ujhelyi 1996:74). This should not be a surprising observation, since the intermediate stage of language origins should, theoretically, lay within the early hominid lineage. According to evolutionary theory, the nonhuman primate and hominid lineages ‘split’ approximately six to eight million years ago (Tattersall and Schwartz 2000). For instance, it has been proposed that chimpanzees have remained physiologically unaltered for the last five million years, while the hominid lineage has undergone at least 17 ‘speciations’ with associated physiological changes during that same period of time (ibid.). All of following lines of inquiry of nonhuman primate communication should be viewed as possible modes by which later hominids incorporated into early linguistic forms.

The homologies between language and animal communication have been questioned by a variety of researchers (Burling 1993; Tomasello and Call 1997; Cheney and Seyfarth 1990a). It is generally agreed that most aspects of human vocal production are shared with a variety of other animals, thus it can be examined from the perspective of comparative evolution (Fitch 2000:258). One can hypothesize that the ability to vary the arrangement of vocal elements, and to produce meaning differences, may be the source of language origins (Ujhelyi 1996:71).

There are three main problems with research concerning nonhuman primate communication: 1) it has tended to focus upon the following specific types of communication: alarm calls, long (distance) calls, response calls, and duetting; 2) very few nonhuman primates have been intensively studied (i.e. vervet monkeys, diana monkeys, macaques, chimpanzees, gibbons); and 3) there is a sharp division between the interpretation of observational field reports of wild nonhuman primates, and the analysis of playback experimental studies conducted upon both wild and captive nonhuman primates. Specifically, the anecdotal evidence seems to imply that nonhuman primates have a rich social life inclusive of specific referential communication among conspecifics, whereas playback studies point to simpler, behavioural mechanisms.