Human Language: A theory, some evidence, and a bonobo to muck it all up again

    Professor Grobstein pulled the photo of a white, oblong, something down onto the screen. It’s certainly is something; a long white cord with a bulbous swelling at one end. Black lines point at areas of potential interest without explaining what they are. This, our Professor explains, is an input-output box, and it is composed of many upon many other smaller input-output boxes. What’s more, we all carry something analogous to it between our ears day in and day out. Here’s the diencephalon, here’s the spinal cord; we’re looking at the central nervous system of a frog.(1) The Professor then asks us, given that we share so much with this invertebrate nervous system, what is it that makes us different? Anatomically, the neocortex seems like a satisfactory answer, but this surplus of dark matter winding its way into gyri and sucli is a mammalian trait. Beyond that, further, what else makes us different? What do we have or do that other organisms don’t? The only answer I could offer was more words. Human beings are the only beings who can engage in a discourse with themselves, with others, immediately, over time, about things present and about things which will never exist. We are a species defined by our verbal behavior.

    Yet what is language? Like many characteristics thought to distinguish one group of highly varied organisms from another, there is a great deal of discord concerning what exactly it is that separates us from “lower” animals and how to go about discussing it. The general consensus is that language is the ability to refer to objects or events, real or abstract.(2) It is voluntary, as opposed to elicited, and a speaker combines pieces of it to generate new concepts.(3) Language is inextricably bound into society and culture, but unlike other social behaviors it is not overtly taught.(4)  It doesn’t seem sensitive to items we traditionally consider rewarding(5) and although it follows predictable rules, it doesn’t seem to fit cleanly into any box theorists have drawn up for it. While other papers will take the time to split the three concepts separately, during this conversation language, speech and verbal behavior will be considered equivalent. There is disagreement in the field as to whether or not language is a completely novel characteristic of man, with no precursors to be found in the behavior of his ancestors or if the use of language is an exaptation of another cognitive, emotional or communication process. Whether or not animals use a system of communication which resembles or lead to our own, language as humans use it is not be found anywhere else on the planet.

    Given how controversial pining down a definition for language can prove to be it should come as no surprise that the origin of verbal behavior is also a somewhat contentious topic: in the past gods and spirits were called upon for an explanation, while more recently a researcher could appeal to a biologist, psychologist, linguist, anthropologist or any combination thereof.  I will present an account put forth by B. F. Skinner in his piece “Selection by Consequences.”(5) In Skinner’s beginning there was behavior, simply exchanges between an organism and its environment. Operant behavior, however, is a specific kind of behavior wherein the consequences of an action are said to influence a subsequent action, i.e. behavior selected on the basis of its consequences. Behavior selection by consequences can be observed on any of three levels of analysis: the phylogenic, which corresponds to natural selection on an evolutionary scale, the ontogenic, which refers to shaping within an organism’s lifetime, and finally on the cultural level, where selection acts upon the behaviors of a group. According to Skinner, the advent of verbal behavior and human society coincided with the achievement of operant control over our vocal musculature. That is to say we were only able to become social when our verbal utterances became sensitive to their consequences in lieu of being determined by phylogeny.(5) Once operant control of the verbal apparatus had been established, the fodder which culture acts upon is produced and cultural selection can occur. Where Skinner left off discussing the relationship between operant control and cultural selection Charles A. Catania picked up in his book chapter “The Natural and Artificial Selection of Behavior.”

    Apes, like other social mammals, have a relatively complex repertoire of vocalizations at their disposal to address the need for communication within a group, so it is not unreasonable to assume, if only for the time being, that our pre-verbal common ancestors shared these skill sets as well. The initial function of these vocalizations was to somehow influence or change another organism’s behavior, be it to get another primate to bring you food or to warn your offspring of impending danger. They were sensitive to the environment in that certain events could trigger them, but voluntary production did not yet exist. The leader of this hypothetical pre-verbal group produces large proportion of these calls to which the other members of the pack respond. Over time, leaders who made more adaptive calls for their group were able to maintain better control over their groups. Adaptive calls were passed on in phylogenic terms, however as the environment changed, leaders sensitive to the changes were able to produce novel and adaptive calls, and were more successful.  Even though leader vocabularies increased in response to the environment, these changes remained within the animal’s lifetime. Their personal calls died out with them until an animal capable of imitation evolved. A capacity for vocal sound rehearsal is critical to the development of language and culture, but imitation is no easy task. Not only does the vocal apparatus need to come under operant control, but auditory discrimination needs to be precise enough to provide something to copy. Eventually a leader was able to pass on his calls to another ape who continued to produce those utterances as long as they were supported by the context. When a vocalization outlived its creator, cultural selection was born. Behaviors were no longer selected based on their heritability, but rather on the ability for the behavior to self-replicate. The ability to imitate and rehearse calls from those gone past also demanded that the new proto-verbal apes learn to discriminate between the calls from their leader, the calls they repeated to themselves in his absence, and finally the calls which came form “within.” Discriminating one’s own voice is the beginning of self-awareness.(4) & (5)

    In his account of verbal behavior, Skinner argues that the operant control necessary for producing language “...resulted from the evolution of a special innervation of the vocal musculature...” This conjecture provides a researcher with something to look for, a change in the fossil record indicating that proto-humans were developing control of their voice-boxes. The problem with this task is that language is formed by body parts which do not fossilize. Instead of searching for a “smoking-larynx,” researchers have begun searching for other indicators increasing operant control would leave behind. For humans to produce speech, our breathing, our mouths, tongues and lips, must be subject to voluntary control. The human family tree is rife with gaps and questions marks (7), so it’s hard to assert that any specific change occurred at any specific time. Given this caveat on the data, comparing the size of various foramina which once housed nerves responsible for the control of various structures has indicated when the onset of language may have occurred. The first hole to consider is the hypoglossal canal, a foramen near the base of the skull which permits the hypoglossal nerve to innervate the tongue. Modern humans and Neanderthals have significantly larger hypoglossal nerves than the older Australopithecus and Homo halibis, suggesting that speech sounds could have begun around 4000,000 years ago. A second bony hole, the thoracic vertebral canal, houses the phrenic nerve which is the primary motor neuron of the diaphragm and is responsible for our voluntary control of breathing. Thoracic vertebral canals approached modern sizes sometime between
1.6 million to 100,000 years ago, meaning the capacity to control breath is probably older than the articulation of the tongue.

    A few weeks after Professor Grobstein asked us to consider what it was that made humans different, Thomas Nagel’s essay “What is it like to be a bat?”(6) found its way into our class discussion forum. Attempting to tease out a distinction between subjective and objective, Nagel proposes that we humans are limited by our own perception systems to our own experiences and that we can never truly know what it’s like to be any other animal. To work within his metaphor for a bit, because we humans have no sonographic sensibilities we cannot fathom the subjective experience of being a bat, which relates to its world primarily in this medium. If language is a characteristic behavior which ditinguishes humans, much as a bat’s sonar isolates it from other mammals, then what kind of an experience is the bat denied because it cannot produce speech? More importantly what, if any, limits does being verbal creatures place on us? Can we even imagine existence without language? Can we think without words? Surely the first brains to produce language were very different from our own, but what changed? Were our brains themselves specifically re-wired over time for the sole purpose of producing language? Are our minds slaves to speech production?

    Skinner and Catania assert that self-awareness is just verbal behavior turned inward to private events, restricting consciousness to us hairless apes with exquisite diaphragm control. After all, what good is a private event or sense of self without the capacity to relate it to another?(4) Years of American Sign Language experiments with apes and chimps seem to show that even our closest living relatives are incapable of developing the advanced grammar rules necessary for reliable, self referential discourse.(2) Nagel, however, would not be satisfied by such an account. The failure for primates to produce verbal behavior as we do is likely related to our limitations as highly verbal creatures. Although speech it is still a learned behavior, portions of its development are determined by phylogenetic processes which prepare us to learn. We don’t need to explicitly teach language to our young because humans evolved to imitate the speakers around us. Our vocalizations are sensitive to the consequences arranged by our social group. (3) Perhaps an ape failing to speak or learn sign-language is not so much a short coming of the student, but of the teacher. Because we don’t know what it’s like to have a pre-verbal brain, perhaps we don’t know how to teach language to a brain without our language learning strategy.

    One would think, however, that given the amount of time and research that went into the ape language studies that at least one ape and one research team would reach the linguistic goal. Some readers may have already thought of Koko the gorilla as an example of primate language acquisition, but Koko’s case is tricky. First she was explicitly taught over a very long time how to “talk” and her signs often remain ambiguous meaning her handlers must translate for Koko, providing plenty of opportunity for biased interpretations. (8) While Koko’s primary handler Dr. Patterson has not been forthcoming with her research methods in the scientific community, she has made sure that the general public is aware of the gorilla. These ape studies argue against Nagel within the framework of Catania and Skinner: because there is no verbal behavior, there is no room for animal consciousness. Susan Savage-Rumbaugh and Duane Rumbaugh’s work with bonobos, small relatives of chimpanzees, also appeared to be supporting the view that language was limited to humans. The researchers were unsuccessfully trying to get a female, Matata, to use a keyboard of symbols to communicate in an attempt to reduce the ambiguity of signing and eliminate the potential for the animal to cue off the trainer (a Clever Hans error). While Matata was explicitly taught, her infant foster son Kanzi was ignored. That is, until Kanzi began using the keyboard himself. Kanzi was the first recorded non-human to show early, spontaneous symbolic representation. The precocious bonobo also appears to be able to comprehend simple spoken English. When tested with syntactically correct but bizarre and novel requests, Kanzi performed at a level comparable to that of a two-year old. He produced and follows two simple rules of syntax to impart additional meaning by word order. Kanzi’s sentences are often short, but unlike other simian communications, his requests often concern two other humans, not himself. Savage-Rumbaugh and Brakke have since developed a paradigm in which a bonobo and a common chimp are raised together for four years. Whereas previous attempts to produce language comprehension in common chimps had consistently failed, the new assisted set-up produced two animals capable of “basic language acquisition.”

    The ability for a bonobo to understand language seems to indicate that the neural substrates for comprehension may be much older than the operant changes required for speech production. (9) And while there doesn’t seem to be any indication that Kanzi has turned his dialogue inward and is the first documented "self-aware" non-human, Skinner and Catania’s model cannot directly account for an earlier development of language comprehension. Neurological surveys of macaque brains also seem to indicate that the region homologous to a human Wernicke’s area, a region of the brain thought to be related to language comprehension, is older than the other language processing centers.(2) That language comprehension seems to be occurring in “lesser” primates detracts from the popular hypothesis that language was a completely new and spontaneous accomplishment of human evolution. Instead  it would appear that our behavioral trademark is the result of selective forces acting through phylogeny, ontogeny and culture to produce our  most interesting way of making noise.  




Sources Cited and Fun Links:

1.  Of frogs and humans.... and birds
/bb/neuro/neuro05/neocortex.html

2. Aboitiz, F. & Garcıa, R.V., “The evolutionary origin of the language areas in the human brain. A neuroanatomical perspective.”in Brain Research Reviews, Vol. 25, 381-396, 1997.

3. de Boer, B. “The Evolution of Speech,” An Excerpt from Encyclopedia of Language and Linguistics, Elsevier http://www.ai.rug.nl/~bart/deBoerEncyclopedia.pdf

4. Catania, C.A. “The Natural and Artificial Selection of Verbal Behavior” in Behavior analysis of language and cognition

5. Skinner, B.F. “Selection by Consequences” in Science, Vol. 213(4507), 501-504, 1981.

6. Nagel, T. “What is it like to be a bat?” http://members.aol.com/NeoNoetics/Nagel_Bat.html

7. “Early Human Phylogeny,” at Human Evolution at the Smithsonian Institute
http://anthropology.si.edu/humanorigins/ha/a_tree.html

8. “Are gorillas using sign language really communicating with humans?” at The Straight Dope
http://www.straightdope.com/columns/030328.html

9. Lloyd, E. A. “Kanzi, evolution and language,” in Biology and Philosophy, Vol. 19, 577-588, 2004.

Video of Kanzi Using his Keyboard
http://www.youtube.com/watch?v=kwm4FEB9LC8

Article about Kanzi & New Research in Vocal Communication
http://www.smithsonianmag.com/science-nature/speakingbonobo.html

Kanzi’s Website at the Great Ape Trust of Iowa
http://www.iowagreatapes.org/bonobo/meet/kanzi.php