If we could talk to the animals, learn their languages

Think of all the things we could discuss…

(From the musical, “Dr. Doolittle”)

More than a century ago, in his chapter on language in “The Mental Evolution in Man,” George Romanes [1] wrote that, etymologically speaking, language is sign-making by the tongue. He provided a schematic hierarchy of language or sign-making; and, except for that which he considered “intellectual,” all such “faculties” were common to humans and nonhuman animals. He cites anecdotal evidence of language among wasps, bees, ants, birds, and other mammals, with “vocal tones being used as intentionally significant of states of feeling and of definite ideas…” [p. 96]. Romanes [1] continues citing other instances of communication by ponies, elephants, dogs, and monkeys. At the end of the chapter, Romanes [1] cites experiments by John Lubbock on “teaching animals to converse” [p. 101].

Romanes was a friend of Charles Darwin and an ardent advocate of his theory of natural selection. Darwin’s Expression of the Emotions in Man and Animals [2] was published in 1872, little more than a decade earlier. In it, Darwin stated that “not only has the body been inherited by animal ancestors, but there is continuity in respect to mind between animals and humans” [2]. As one of Darwin’s most enthusiastic supporters, Romanes collected many anecdotes about animal behavior which he likened to humans. As noted in the above paragraph, Romanes went so far as to characterize language in nonhuman species.

Speech allows man to utter what he does not think.

(Attributed to Thomas Hobbes)

Language is considerably more wide ranging than sign-making by the tongue and includes verbal and nonverbal communication: spoken language—speech—that involves coordinated motor actions utilizing orofacial musculature as well as other muscles in the vocal tract, nonverbal gestures which are discernible bodily actions that impart meanings, or writing of any kind. From an ontological perspective, preverbal infants communicate by gesturing or pointing. All known living human groups have language; and, certain features of the human body, such as the auditory channel and vocal tract, fit the needs of speech. Linguistic ability has been linked to regions in the brain [3]. According to Pinker [4], these biological facts suggest that human language is a natural outgrowth of human physiology. In accord with Darwin’s thinking, and irrespective of the concerns raised by Botha [5] regarding the specificity and explanatory power of the Pinker and Bloom argument, Pinker and Bloom [3] consider language and language acquisition in humans as the product of evolution and natural selection, and not an innate grammatical ability.

On the other hand, there are linguists who regard language as an innate function contained in modules residing in the brain [6, 7]. These mental faculties are organized in such a way that acquisition and elicitation of language is the outcome of interactions among the several modules. For almost half a century, Chomsky has argued that linguistics is a hypothetical construct, not a biological fact, and therefore not a subject of scientific inquiry. Specifically, Chomsky [8] states: “… language and higher mental faculties generally are not part of biology” [8] [p. 62]. He echoes remarks by Hauser that “language is not properly regarded as a system of communication. It is a system for expressing thought” [8].

Chomsky adopts the point of view espoused by Mountcastle that “things mental … are emergent properties of brains.” Moreover, according to Chomsky, when studying communication among animals, both human and nonhuman, one needs to embrace four principles:

Herein lies a dilemma: One could continue to pursue in greater detail Chomsky’s Cartesian point of view regarding emergent properties from innate “language acquisition devices” (LAD), attempt to rescue human and nonhuman communication within the principles outlined by Chomsky, then compare and contrast them to the development of language from a biological and evolutionary standpoint, as set forth by Pinker and Bloom and others. There are notable strengths and weaknesses to be found in each of these models. However, since the purpose of this chapter is to examine the characteristics of speech and language in human and nonhuman animals and how they may be related, I will focus primarily on and attempt to recapitulate the biological–evolutionary argument for human and nonhuman communication and how human language may have evolved from nonvocal communication.

At the heart of Chomsky’s model is an innate mechanism located somewhere in the brain, an LAD that employs a universal grammar (UG), to account for the variety and rapid acquisition of language across cultures. A century earlier, Darwin [9] had also noticed differences in language formation but saw a parallel between linguistic change and species change as evidence of his theory of evolution. In a similar albeit more modern vein, Pinker and Bloom [3] describe language as no different from any other multifaceted ability in animals, for example, echolocation in bats; and, the only scientific means by which to explain the development of these skills is through evolution and natural selection. Moreover, although the rapid acquisition of language would appear to favor Chomsky’s thesis, Christiansen and Chater [10] propose that language has evolved to be something learnable. That is, language is wrought by the brain and is a function of the neurobiological and neurophysiological mechanisms residing in the brain extant for learning.

Anecdotal precedents of a mouth-gesture theory have been dated from an essay by Condillac in the mid-eighteenth century [11]. However, it was Alfred Wallace—the co-progenitor of the theory of evolution with Charles Darwin—who, in 1881, first formally proposed a mouth-gesture theory as the basis for the emergence of speech in humans [12]. Wallace’s model was later endorsed by L. H. Morgan, the founder of anthropology in the USA. In the twentieth century, one of the earliest systematic discussions concerning a relationship between expressive speech and hand gesture appears in a lecture by Brain [13]. Brain asserted that “the dominance of one hemisphere was a precondition of symbolic thinking and expression which at the psychophysiological level distinguishes man from all other animals, and we may guess that when human culture first developed man was already both able to speak and predominantly right-handed” (1949; p. 840). Hewes [11] also remarked that evolutionary selection pressure may account for the unusual link between left-hemisphere dominance and right-handedness.

From a biological point of view, the model proposed by Hewes [11] positing a shift from manual gesture to vocal language for humans would be a relatively uncomplicated one to arrange, despite the fact that a vocal as opposed to gestural language system requires modifications in the lips, tongue, larynx, and other structures along the vocal tract. However, such a modification would not have been possible for nonhuman primates since their vocalizations are primarily set in motion by social stimuli and not elicited autonomously. Moreover, the ability to assimilate visual–vocal signals requires the ability transfer learning across modalities, a skill monkeys do not demonstrate when spoken to by humans. Hewes [11] also noted that a transformation from gestural to vocal language has several evolutionary advantages. Specifically, auditory signals are more readily detected than visual cues in dense, foggy, or poorly lit environments and can be received over greater distances. Moreover, hand gestures require more time and energy to transmit.

Corbalis [14] recommenced the argument put forth by Brain [13] that, irrespective of nationality or culture, humans are mostly right-handed, and nonhuman primates typically show no such asymmetry, more often favoring the left hand. Right-handedness has been demonstrated repeatedly in every human subpopulation. Still, the lack of right-handed asymmetry in nonhuman primates has not always been corroborated. Meguerditchian et al. [15] assessed 162 baboons for handedness with respect to gestural communication and found right-handedness in significantly greater evidence. Meunier et al. [16] examined hand preferences in human infants and adult baboons and found that hand preferences for pointing—but not grasping—in both humans and baboons significantly favored the right hand. As noted earlier, right-hand dominance is typically thought to be a contralateral function produced by the left-hemisphere region in the brain. To examine the relationship between handedness and hemisphere, Knecht et al. [17] used the Edinburgh handedness inventory and verified that left-hemisphere dominance was strongly associated with right-handedness. Conversely, only 27 % of extreme left-handed human participants displayed right-hemisphere dominance. In other words, more than 70 % of extreme left-handed individuals also exhibited left-hemisphere dominance.

It is also the case that the left hemisphere contains regions in the brain—Broca’s and Wernicke’s area—which have been identified with speech, although Lieberman [18] has argued that language loss from lesions in these sectors does not occur without additional subcortical damage. Previously, researchers found that left-hemisphere asymmetry is associated with vocalization in infrahuman species from monkeys [19] to frogs [20]. Aboitiz [21] proposed an alternate hypothesis to that of Brain’s [13], i.e., hemispheric dominance for language may be more efficiently encoded in one hemisphere, which would agree with the lateralization found in songbird song production [22], a matter to which we will return later.

A more complex picture emerged when Skipper et al. [23] reported that neurons in primates’ brain regions associated with visual, auditory, and motor activity are triggered when they are performing an action or when they perceive a conspecific performing an action. In their study, Skipper et al. [23] found that audiovisual language comprehension generated activity in regions associated with sensorimotor aspects of speech production in humans. More recently, Skipper et al. [24] recruited 12 right-handed young adults and examined five regions of interest in the brain associated with motor areas implicated in speech perception, gestures associated with speech production, and areas involved with language comprehension. They found that brain regions linked to motor activity production were also associated with gestures and facial movements related to language comprehension, affirming Lieberman’s [18] earlier remarks regarding the link between Broca’s area and the subcortex.

The motor theory of speech perception made a great leap forward when “mirror neurons” were discovered in the monkey’s ventral premotor cortex, area F5; although, more than a decade earlier, Williams and Nottebohm [25] had found that motor neurons in zebra finches responded to auditory stimuli. F5 in monkeys is considered analogous to Broca’s area in humans [26]. Other brain regions in monkey, such as the superior temporal sulcus (STS), also contain neurons that respond to biological activity. Both F5 and STS are linked to the inferior parietal lobe, area PF, and are referred to as “PF neurons” [27]. Additional areas in the brain, e.g., the amygdala and orbitofrontal cortex, which are involved in social and emotional comprehension, are also thought to be part of the mirror neuron network. Thus, both language development and social behavior appear to be related to a particular type of neuron and located in specific interconnected brain regions. Additional evidence of commonalities in human and nonhuman primate brain was found by Petrides and Pandya [28], who examined areas 44 and 45 in the macaque monkey, two separate but adjacent regions in the left-hemisphere ventrolateral lobe, which are homologous to Broca’s area in humans. Petrides and Pandya [28] found comparable neural pathways from these areas to the posterior parietal and temporal regions in both species. Iriki and Taoka [29] hypothesized that the neural mechanisms in these regions, which subserve tool use, may link gesture and language in a way that utilizes the same systems for spatial information processing. In that regard, Macellini et al. [30] were able to demonstrate that macaques were capable of learning how to use tools and in the appropriate settings. Macaques then became sufficiently adept at using tools to be able to employ them in novel situations.

As a case in point that links tool use, handedness, gesture, and language into a distinctive, well-defined configuration, Calvin [31] examined the throwing precision of primates, noting that the timing and accuracy of objects tossed by primates at other animals—or hammering nuts with rocks to break them open—could not be accidental; it required great accuracy and was likely relating to “timing neurons” and increased brain size. Moreover, Calvin [31] observed that the overhand motions implicated primate rock tossing was not dissimilar to that of hominid spear throwing, either of which could be used as a gesture to ward off predators or with a weapon to kill game.

However, despite the similarities in human and nonhuman primate brains and handedness asymmetries, nonhuman primates have never demonstrated proper language capability, nor have they been able to acquire proficiency in tool use by watching another skilled individual utilizing a tool [30]. While vocal communication other than gesture among animals is ubiquitous, nonhuman primates evince little in the way of its regulation via brain activity. Corbalis [32] noted that animals employ vocalizations to draw attention to events about which other conspecifics may be unaware, e.g., evidence of predators or food in the area or for mating or parent–offspring interactions. However, if linguistic ability is comprised of a common enterprise involving joint attention between conspecifics, the question remains as to whether nonhuman primates, while capable of making gestures and producing vocalizations, use either as a form of communication with conspecifics other than to denote danger, food, desire for sex, or parenting.

Tomasello and his colleagues studied several species of nonhuman primates to determine if their gestures could be ascribed to nonverbal communication. An early goal was to catalog gestures produced by one species of great ape, the gorilla [33]. Pika et al. [33] observed young gorillas and adults daily in two settings over a 7-month period and found that gestures performed by adults were acquired by young gorillas at an early age (1–2 years). Liebal et al. [34] examined 19 chimpanzees 3–4 times a week over a 6-month interval and noted that most visual gestures were made only a when a conspecific was visually attending, suggesting that chimpanzees are aware of what conspecifics can and cannot see. On the other hand, Povinelli and Eddy [35] found that chimpanzees would gesture for food from a human even when the human was blindfolded or had his head covered, thereby indicating that chimps may not always have a concept of visual perception in others. Although gestures, particularly gesture sequences, were elicited in several contexts, e.g., competition, nursing, and sexual behavior among them, the largest proportion of such movements tended to be gesture sequences involving play. In an earlier study of chimpanzees, Tomasello et al. [36] examined learning processes used by young chimpanzees and found that they acquire gestures early on primarily by “conventionalization,” i.e., signals created between two conspecifics when they attempt to communicate with one another. Pika et al. [37] later referred to this type of interaction as “ontogenetic ritualization.” Despite their extensive assessments, Tomasello et al. [36] found no evidence of imitative learning, a social process. Also, Tomasello et al. [36] did not find a strong association between signals and contexts: Some chimpanzees used a well-defined signal for many different purposes, other chimpanzees used a variety of signals for a single purpose. That is, there was no indication that gestural communication among chimpanzees resembled any form of grammar. Similarly, Pika et al. [37] observed bonobos during a 3-month interval to catalog their gestures and determine how they acquire and use them. Consistent with their earlier studies, Pika et al. [37] found no evidence of social learning processes among bonobos and that their chief form of learning was through ontogenetic ritualization.

Another notable feature of linguistic aptitude and communication involves the ability to refer to objects or persons not present, i.e., displaced reference, a skill that presumably differentiates humans from nonhuman primates. To examine this premise, Liskowski et al. [38] tested 16 prelinguistic infants, 12 months old, and compared them to 16 adult chimpanzees in two situations that involved a desired object that no subjects could see. There were two settings for the experiment: (1) an occluded-referent condition in which the desired object—a ball or block for infants, food for chimpanzees—was placed out of sight, under a platform and (2) an absent-referent condition in which the object was removed entirely from its original location. Results showed that more than half the sample of infants could point to the target location—the platform—in either condition and without being prompted, whereas, while half the sample of chimpanzees could gesture to the target location for the occluded-referent condition, no chimpanzee could point to the platform during the absent-referent condition. Liskowski et al. [38] concluded that humans are thus able to communicate about displaced referents, even before language ability emerges, whereas this ability is not present in nonhuman primates.

In any event, if language evolved from vocalization and gesture, whence had it occurred and might it have been acquired?

Interestingly, recent archeological evidence indicated that Homo sapiens and Neanderthals may have had more in common than originally thought, particularly with regard to speech and language. Uomini [39] ascertained that right-handedness could be detected in the remains of Neanderthals. Barney et al. [40] found that vocal tract anatomy in Neanderthal skulls compared favorably to that of modern humans, and the two hominids were not significantly different from one another, whereas the skull and mouth of Homo erectus and other early hominids were significantly different. Therefore, the data suggest that Neanderthals had the anatomical capability to articulate sounds beyond that of simple vocalizations observed in nonhuman primates, and likely to have been the case with our earlier ancestors. Another important finding from Neanderthals’ remains has been DNA evidence which indicates the presence of the FOXP2 gene [41], a gene that has been found recently to be involved with the development of speech and language in humans.

From their human twin studies, Bishop et al. [42] found evidence of a genetic basis for specific speech and language impairment (SLI). However, establishing a genotype–phenotype relationship between a genetic mutation and disordered speech and language has been problematic, owing to the relatively small number of affected families, and could have been made more so if more than one gene was involved in the behavioral phenotype. By using linkage analyses in families with affected members, regions of interest on several chromosomes have been associated with SLI [43]. The first important pedigree identified is known as the KE family, originally thought to have a genetic disorder that primarily disrupted grammar ability [44, 45]. Fisher et al. [46] performed a linkage analysis on the family and surmised that SLI resulted from a mutation of a single autosomal dominant gene. Moreover, based on neurological assessments of affected and unaffected family members, the mutation was also found associated with abnormalities in motor areas in the brain. That is, the gene affected areas in the brain other than those typically associated with grammatical ability. Shortly afterward, Lai et al. [47] examined other features of SLI in the KE family and observed that their severe speech impediments were related to their impaired ability to perform sequenced movements that involved orofacial musculature. This difficulty also impinged on several aspects of linguistic aptitude: the ability to separate words into their respective phonemes, the ability to parse phrases and sentences properly, and the facility to understand other aspects of grammar and syntax. From their molecular genetic analysis, Lai et al. found a mutation in a forkhead-domain gene, FOXP2, located on chromosome 7q31.1, in affected members of the family. As a result, they concluded that the FOXP2 gene was involved in the development of speech and language. A subsequent study by MacDermot et al. [48] of 49 children with a primary diagnosis of verbal dyspraxia found variants in the FOXP2 gene that altered the FOXP2 protein in three individuals.

FOX proteins constitute a relatively large family of proteins which act as transcription regulators. Within the Foxhead family are many subgroups, one of which is FOXP. The FOXP2 gene is one of four members of FOXP. Genes within the FOXP subgroup perform a variety of tasks, mutations in which have been shown to produce developmental disorders [43].