The Evolution of the Human Voice and Speech: Key Components in the Story of Our UniquenessCHAPTER 1

Since we, Homo sapiens, arrived, however, we have become the unquestioned masters of this world. Yes, it was tricky at first, as we possessed neither the great strength of mammoths nor the speed of gazelles. We lacked distinctive dentition, such as massive canines, that could frighten away competitors. We had no claws or even a quill or two. We could not swim very well. We could not climb very well. Neither our vision nor our hearing was particularly impressive. Physically, by mammalian standards, we just “were not all that”, as our kids would say today.

We did accrue, however, something very special along the path of our evolution: a remarkable ability that no other animal on our planet either now or in our long history has had. We have a unique voice – and the ability to use that voice to produce articulate speech. Together, this duo has enabled the panoply of our multilinguistic abilities and has given us our unique, species-specific, language. With these, we have dominated our planet.

Voice, sometimes called “vocalization”, usually refers to sounds that are produced at the laryngeal vocal folds, mostly through air from the lungs. In humans, vocalizations comprise the fundamental components of speech (vide infra), but not all vocal sounds are part of the speech spectrum. Indeed, we utter many involuntary sounds—from coughs to an infant's babble—that are generated at the vocal folds but not morphed into articulated sounds of speech. “Phonation” is the term that describes the production of the voice via vocal fold modifications. The essence of voice is, thus, its tie to vocal fold anatomy, physiology, and neuromuscular control. Of course, “voice” has also come to have a variety of societal, literary, and even philosophical meanings that go far beyond the confines of the larynx.¹4

Speech is more complex than voice to define. For humans, we can consider speech as the verbal–vocal communication system regularly used by all living people. By this definition nonhuman species would thus not have speech. As many species arguably have advanced vocal systems, they are often referred to as having “speech”, e.g. “ape-speech” or “the speech of prehumans”. Accordingly, what we produce is often given the modifier, “human”, and our speech indicated as “human speech” to distinguish what we do from other vocal utterances. For our purposes here, speech will be defined as what we do, and other species vocal communication considered and referred to as just that, vocal communication. Whales and apes may call to complain (please leave a vocal message if you can).

Returning to our species, speech is the product of both the central (brain) and peripheral nervous systems and the aerodigestive tract (ADT), particularly that component of it known as the “vocal tract”. Of particular interest to the otolaryngologist is the supralaryngeal component of the vocal tract, the region that is responsible for the modification of the fundamental sounds (voice) produced at the vocal folds. As will be emphasized below, it is important to always keep in mind that there is not an anatomically or physiologically separate “vocal tract” that functions solely for this purpose, rather the components are inherent parts of a complex, multifunctional ADT that evolved for primary purposes other than speech (a point often overlooked by our linguistic colleagues).

Lastly, there is language, the most complex concept of them all.^2,3 To define language in a few words would do it injustice, but as a working definition for us we can view it as the global expression of human communication either spoken, written, or gestural, consisting of words in a structured, ordered, and conventional manner. Inherent components of human language include brain (mental) functions of cognition, grammar, and syntax, among others. The crucial peripheral components of the input systems (i.e. auditory, visual) and output (voice, speech) are, obviously, inextricable elements.

The anatomical mechanisms for these controls and protections are the internal membranes, their folds, and controlling intrinsic muscles. While a detailed histology of the folds is beyond the scope of this essay, the basic composition of the structure has been divided by Hirano⁹ into five layers: squamous epithelium, superficial lamina propria (SLP), intermediate lamina propria (ILP), deep lamina propria (DLP), and vocalis muscle. These layers essentially function in mechanically “decoupled” groupings of the layers to form: the “Cover” or mucosa (epithelium and SLP); the “Transition area”, the vocal ligament itself (ILP and DLP); and the “Body” of the fold produced by the vocalis muscle.

Although comparative and descriptive histology of the larynx, in general, and vocal folds, in particular, have5 been done, many aspects still remain unclear, particularly as concerns the direct relationship to functions.⁶ Indeed, the relatively new field of “Voice Science”, and how specific structures/layers of the folds directly relate to normal or abnormal function, and medical or surgical treatment is in its nascent stages. Furthermore, what remains largely unstudied and unknown is how the different compartments of the vocal folds morphed and changed over the millennia to accommodate the evolutionary forces compelling their protective and pressure control roles.

While the initial roles of the larynx and its internal membranes of mammals, in general, and primates (prosimians, monkeys, apes, humans, and all of their ancestors), in particular, were sphincteric and protective, phonatory importance may well have played a part, though to what extent is unclear. Certainly, as we came to a hominid (direct human ancestors) grade of evolution, phonation/vocal communication, and even singing (as an important part of social communication) would have begun to take on increasing importance. Such importance would have had a selective advantage in evolutionary terms, working to select for vocal fold histology in ways conducive to sound production, distinct, and with possibly greater ranges than our closest relatives. Some studies are already pointing to this. For example, only humans seem to possess a deep layer of the lamina propria, which may account for species-level differences between humans and our nonhuman primate relatives.¹⁰ Along similar lines, it has been shown that the human thyroarytenoid muscle (the muscle within the body of the vocal folds; the most medial fibers are often named the “vocalis”) may uniquely exhibit slow tonic muscle fibers with rare contraction properties, such as contractions that are prolonged, stable, precisely controlled, and fatigue resistant. Such properties have been suggested to be a unique specialization (called an “autapomorphy”, or uniquely derived characteristic, in anthropological terms) that may underlie our unique “voice” and help effectuate speech.¹¹

It should be noted that while our field is collectively trying to identify that which is special or comparatively distinct about the anatomy/histology of human vocal folds, there remains much we do not know about developmental changes in this regard. Indeed, we humans go through a remarkable developmental scenario, particularly in the changes to the aerodigestive region (vide infra). Corresponding, and perhaps cotemporaneous, developmental changes may well occur in the internal laryngeal environment. Indeed, recent studies have shown that significant histological differences may exist in vocal fold development, for example, the deep layer of the lamina propria is not present at birth and not fully developed until the age of 11 or 12 years.^10,12 While the functional and clinical implications of such observations are not yet fully clear, these may well have major implications for laryngeal functions, ranging from protection to voice production and speech attainment.

The high position of the larynx enables the epiglottis to pass upward behind the soft palate and “lock” the larynx directly into the nasopharynx. This configuration provides a direct air channel from the external nares through the nasal cavities, nasopharynx, larynx, and trachea to the lungs. Liquids, and in some species even chewed or solid material, can pass on either side of the interlocked larynx and nasopharynx by way of the isthmus faucium, through the piriform sinuses to the esophagus, following the so-called “lateral food channels.”6

While the larynx is consistently high in most mammals, its exact position and the extent of its placement in the nasopharynx can vary considerably among species. Studies of cetaceans (i.e. whales, dolphins, porpoises), for example, have shown that the larynx in some species is positioned so high (rostral) that it is no longer in the neck but rather lies largely within the head.¹⁵ In many odontocetes (toothed whales), the larynx, from the tip of the epiglottis to the lower border of the cricoid cartilage, corresponds to the level of the presphenoidal synchondrosis to the caudal border of the basiocciput. The larynx thus usually never reaches as far caudally as the compressed cervical vertebrae characteristic of these7 mammals. The epiglottis and corniculate cartilages form the anterior portion of an elongated larynx, which is encircled by a strong palatopharyngeal sphincter muscle (homologous to the soft palate and palatopharyngeal arch in humans). This sphincter grips the top of the larynx and keeps its aditus intranarial, thus effectively sealing the respiratory tract from the digestive route. Although exhibiting the basic mammalian pattern of a high larynx, odontocetes appear to have exaggerated it by placing the larynx even higher (or more rostral) than their terrestrial relatives. This extrahigh position ensures these mammals of a larynx fitted snugly into the nasopharynx and indeed may make it habitually intranarial, that is, it is not usually retracted from its position behind the soft palate. This arrangement may allow them to swallow whole fish while communicating with each other or while echolocating to orient themselves in their environment.

Although cetaceans demonstrate an example of larynges that have both migrated cranially and elongated their rostral cartilages, some terrestrial species have larynges that have expanded their caudal components (e.g. thyroid cartilage) so that they appear to extend considerably into the neck. For example, some male artiodactyls (red and fallow deer, Mongolian gazelle) exhibit particularly large larynges that seem to be located more caudally in the neck compared with other related species.¹⁶ However, although the larynges of these animals are elongated, they still retain roughly the same position opposite the cervical vertebrae as most other terrestrial mammals (extending from the basiocciput to C2–C3). Maintenance of this typical mammalian position is due to concomitant elongation of the cervical vertebrae. These animals also exhibit an elongated and elastic velum (red and fallow deer) and an elongated epiglottis (Mongolian gazelle) that appear to assist in epiglottic/palatal contact and, therefore, the maintenance of the “two-tube” system. Thus, while larynges differ considerably in position, the basic, ancestral two-tube configuration is essentially maintained. These animals have modified a basic plan; they have not changed it.

Postmortem dissections, and a range of imaging studies, including cineradiography, computed tomography, and magnetic resonance imaging, of our closest relatives, the nonhuman primates, show that their upper respiratory anatomy is also similar to the general mammalian pattern.^4,17–19 As in other mammals, nonhuman primates exhibit a larynx positioned high in the neck, usually corresponding to the first to third cervical vertebrae. This position allows for epiglottic-soft palate apposition and the possibility of an intranarial larynx, thus providing for a direct airway from the nose to the lungs, whereas the alimentary tract passes around the larynx en route to the esophagus (Figs. 1.2A to C). Cineradiographic studies have confirmed that nonhuman primates exhibit mostly separate respiratory and digestive routes and the ability to breathe and swallow almost simultaneously.^17,20 Because of this configuration, nonhuman primates, like other mammals, appear strongly, if not totally, dependent on nasal breathing. As occurs in many mammals, the connection between the epiglottis and the soft palate can be broken, as the larynx exhibits extensive mobility and can be transiently lowered. This can occur for a number of reasons, including some vocalizations, swallowing certain foods (e.g. a large bolus of meat), or due to disease.

Although this anatomic arrangement may enable almost simultaneous breathing and swallowing, it severely limits the array of sounds an animal can produce. The high position of the larynx means that only a small supralaryngeal portion of the pharynx exists. In turn, only a very reduced area is available to modify the initial sounds generated at the vocal folds. Due to this limitation, most mammals therefore depend primarily on altering the shape of the oral cavity and lips to modify laryngeal sounds. Although some animals can approximate some human speech sounds, they are anatomically incapable of producing the range of sounds necessary for human speech.^21,22

The human ADT, particularly the larynx and its positional relationships to contiguous structures, undergoes dramatic changes during development.

The major morphologic events in embryonic development (0–8 weeks) of the human larynx have been well documented and new data from homeobox genes are shedding further light on the mechanisms of early spatial establishment in the head and neck.^23,248

The attainment of larynx–nasopharynx interlocking is a significant maturational horizon in the development of the aerodigestive region. Establishment of this anatomic relationship allows the creation of essentially separate respiratory and digestive routes that will function as such in the newborn infant (vide infra). Our ultrasound investigations have shown upper respiratory activity patterns that strongly suggest an operational two-tube system is beginning to function prenatally, in which the larynx remains highly positioned and intranarial during fetal swallowing movements. A critical time in the9 development of the entire upper respiratory region may take place during the period between weeks 23 and 25. Not only does the larynx attain a position that places it intranarially, but contiguous portions of the skull base—the de facto roof of the upper respiratory tract (vide infra)— also appear to be undergoing remodeling at this time. As portions of the cranial base are intimately related to the larynx and its contiguous musculature, cranial shape development and laryngeal/ADT positions may also be linked, although the precise extent and relationships are still unclear. What may be beginning during this period is a remodeling and refinement of the positional anatomy of the entire aerodigestive region—soft tissue and cartilaginous structures, such as the larynx, and bony skeletal parameters, such as the skull base—to provide the anatomical framework for the newborn's upper respiratory and digestive tract. It should be noted that while these laryngeal and basicranial modifications are occurring in the upper respiratory region, concomitant changes are also occurring in the lower respiratory tract. For example, the period from weeks 23 to 25 corresponds to the maturation of the pulmonary glandular epithelium. This alveolar epithelium is responsible for the production of fetal lung surfactant, a substance that is essential for independent respiratory function. The contemporaneous development of the fetal larynx to permit soft palate–epiglottic overlap with increasing levels of lung surfactant suggests that the time frame for the maturation of the upper and lower respiratory tracts are closely related. Normal fetal maturation of the larynx may be an essential factor in determining the beginnings of respiratory independence and overall fetal viability.

The morphologic pattern of high laryngeal position established during fetal life continues past the perinatal period and into infancy. Indeed, the newborn/young infant period may more accurately be seen as an extension of the pattern established during the late second and early third fetal trimesters rather than as a distinct entity. Postmortem dissections and imaging studies have shown that the positional relationships in the aerodigestive region of human newborns and young infants closely resemble the basic primate and mammalian pattern. In newborns and infants until approximately 1½ to 2 years of age, the larynx remains high in the neck.^4,14,28–32 Its position corresponds to the level of the basiocciput/C1, extends to the superior border of C4 in newborns, and descends slightly to the level between C2 and C5 by approximately 2 years of age. The tongue at rest can be found entirely within the oral cavity, with no portion of it forming the upper anterior wall of the pharynx.

Largely separate respiratory and digestive pathways, similar to those described in most other terrestrial mammals, are effectuated by the high laryngeal position in newborns and young infants (Figs. 1.3A and B).

Although the high position of the larynx in a human newborn or young infant effectuates the dual-pathway system, it severely limits the array of sounds babies produce. Many studies have shown that the high position of the larynx greatly restricts the supralaryngeal portion of the pharynx/tongue available to modify the initial, or fundamental, sounds produced at the vocal folds. Thus, an individual with a larynx situated high in the neck, as is found in a newborn human or monkey, would have a more restricted range of vocalizations available than would individuals with larynges and tongues placed lower in the neck. Indeed, linguistic analyses have identified the quantal vowels [i], [u], and [a] as sounds that human infants or nonhuman primates cannot produce. As these vowels are the limiting articulations of a vowel triangle that is language universal, their absence considerably restricts speech capabilities.

Although the larynx remains high in the neck until around the second year, functional changes, such as the first occasional instances of oral respiration, have been noted to occur considerably earlier, indeed within the first six months of life.²⁹ The period between four and six months, in particular, may represent a crucial stage in upper respiratory activity. At this time, neuromuscular control mechanisms of the larynx and extralaryngeal pharyngeal control are beginning to change even before true structural “descent” of the larynx has occurred. This changeover period may also indicate a time of potential respiratory instability because of the transition from one respiratory pattern to another.

The time and manifestation of both prenatal and postnatal maturation within the central nervous system relates directly to normal and abnormal upper respiratory and swallowing functions. For example, studies from our laboratory have suggested that there are crucial prenatal periods for the development of mammalian upper respiratory motor nuclei in the brainstem and that insults in utero could affect postnatal functions.^33,34 The combination of subsequent, postnatal central nervous system maturation and developmental changes in respiratory patterns may predispose the infant to several developmentally related problems. The sudden infant death syndrome, SIDS, for example, may be related to these first postnatal upper respiratory changes and to the subtle changes in laryngeal position or central and peripheral neuromotor control of the larynx.^33,35 The precise time of the shifts that occur in breathing patterns, their relationship to laryngeal changes, and the neurophysiologic mechanisms that accompany them are obviously crucial questions that are still poorly understood and require more specific and detailed study.

The larynx of human infants may remain high in the neck until approximately 1½ to 2 years of age. Our studies and others have confirmed that around the second year, children begin to show positional rearrangements of the ADT that differ sharply from the condition in newborns and early infants. Around this period the larynx begins its permanent, structural descent into the neck. Although minor topographic changes in laryngeal position continue until puberty and beyond, the major qualitative change probably occurs between the second to third years of life. Although the exact timing of laryngeal descent has yet to be fully determined, imaging and postmortem observations indicate that by the third year the position of the larynx has been significantly lowered.

Although the internal nature of the larynx changes relatively little after the third year (except for normal maturational changes at puberty), positional changes relative to contiguous upper respiratory structures are considerable. The tongue no longer lies entirely within the oral cavity at rest, as in the newborn infant or nonhuman primate. The posterior portion of the tongue, from the foramen cecum caudally, has descended into the neck and now forms the upper anterior wall of the pharynx. The larynx is now situated considerably lower in the neck. For example, in a seven-year-old child, the larynx—from the tip of the epiglottis to the inferior border of the cricoid cartilage—corresponds to the level between the upper border of C3 and the lower border of C5. In the adult, the larynx has further descended, lying between the lower border of C3/upper part of C4 to the upper border of C7. Concomitantly, the hyoid bone and its associated suprahyoid and infrahyoid muscles are relatively lower in the neck.11

Due to the descent of the larynx, tongue, and hyoid apparatus in children after the second to third year, the epiglottis can no longer approximate the soft palate, even during maximal laryngeal elevation. As the larynx cannot lock into the nasopharynx, there no longer exists the possibility of a continuous, tube-like airway from the external nares to the lungs. The lower position of the larynx alters dramatically the way humans, after the early years of life, breathe and swallow. The loss of the ability of the epiglottis to make contact with the soft palate means that the possibility of having two separate pathways, one for air and one for liquid—the basic ancient terrestrial template—no longer exists. Unlike this general mammalian pattern, the respiratory and digestive tracts now cross each other in the area of the pharynx. This low position of the larynx also results in a large supralaryngeal portion of the pharynx. A permanently enlarged oropharynx is now present even during maximal laryngeal elevation. These changes have pronounced effects upon our respiratory behavior. For example, we are habitual nose breathers but, unlike newborn infants, adults have a greater ability for oral respiration and do so more frequently. Loss of the two-pathway system has now made it imperative that the respiratory tract be sealed off during swallowing.

The permanent intersection of the respiratory and digestive pathways has created a de novo “aerodigestive” tract, a first of its kind in mammals. This has created many problems, which, if not unique to us, are certainly accentuated in our kind. A major problem is that a bolus of food can easily become lodged in the laryngeal aditus leading to the common occurrence of food “going down the wrong pipe”; however, if the bolus is large or not expelled rapidly enough one may choke to death. This event is often referred to as a “cafe coronary”, because it frequently occurs in restaurants and may be mistaken for a heart attack. Similarly, another disadvantage of the crossed pathways is the relative ease with which vomitus can be aspirated into the larynx and trachea and passed to the lungs.

Not only has a permanently lowered larynx proven to be a danger in the ingestion of material, but it may serve as the anatomic basis allowing for gastroesophageal reflux of stomach contents to enter the pharyngeal or oral cavities.^36,37 The human ADT has clearly not been evolutionarily selected for the purpose of efficiently dealing with constant, retrograde emissions into the supraesophagus. Indeed, our ADT appears to be particularly poorly designed to handle any esophageal or gastroesophageal reflux. This is most clearly demonstrated by two features: (1) our uniquely low laryngeal position and (2) the relatively unprotected posterior larynx. Low laryngeal position, by creating a permanent oropharynx, has by definition created a greatly expanded supraesophageal region. It should also be noted that as the larynx has migrated caudally, so too has the location of the cricopharyngeal sphincter (i.e. the upper esophageal sphincter) at the caudal-most extent of the pharynx. Thus by definition, adult humans have a relatively shorter esophagus and relatively longer supraesophagus than most other mammals. At the very least, then, the highly acidic contents of human gastroesophageal reflux have gained access to a proportionally greater surface area of pharyngeal mucosa with which they may interact.

Laryngeal descent in humans has altered considerably the way we breathe and swallow. From a comparative perspective, we have lost the basic mammalian ability to breathe and swallow simultaneously or almost simultaneously. We have also accrued a number of most unwanted guests, including the relative ease of lodging material in the airway or refluxing contents to the supraesophagus and other portals. A litany of other diseases and/or incoordination clinicopathologies, ranging from otitis media and frequent rhinosinusitis³⁸ to SIDS and aspects of obstructive sleep apnea³⁹ may also be a price we pay for peripheral and central neuromuscular rearrangements that accompany ADT modifications. The lowered position of our larynx has, however, provided one major positive aspect: a greatly expanded supralaryngeal portion of the pharynx. This enlarged supralaryngeal pharynx has “liberated” the tongue from the oral cavity confinement, now having its posterior portion form the movable anterior aspect of this enlarged chamber. This configuration allows for enhanced oral tidal respiration, which may have occurred in temperate environments for our early ancestors. In addition, pharyngeal/lingual modification of sounds produced at the vocal folds is considerably greater than that possible for newborns, early infants, or any nonhuman mammal. In essence, it is the unique marked descent of the larynx and the resultant expansion of the pharynx and liberation of the posterior tongue that gives us the anatomic ability to produce fully articulate speech.

While fossilized throats do not remain, some researchers, our lab included, have developed methods to reconstruct the structure of the region through the use of remnant portions of the cranial base, the de facto “roof” of the ADT, as a guide.^40–43 Reconstruction of the ADT of fossil human ancestors – ranging in age from our earliest direct relatives known as australopiths (who can be traced to over four million years before the present) to early members of our own genus, Homo (appearing over a million and a half years ago), to our own species, Homo sapiens (arriving perhaps 200,000 to 300,000 years before the present)—have enabled us to trace the changes in the region through our history. For example, our reconstructions have suggested that the earliest hominids likely exhibited an ADT largely similar to those of the extant apes, with the larynx positioned high in the throat and the epiglottis able to contact the soft palate during normal tidal respiration. These early ancestors likely breathed and swallowed essentially as do our living monkey and ape relatives, being essentially nasal breathers with a modified two-tube system. The high position of their larynx would, by necessity, limit the supralaryngeal area of the pharynx and the freedom of the tongue to modify sounds as extensively as modern adult humans can. This suggests that they were restricted in the types of sounds that they could make, probably being incapable of producing a number of the universal vowel sounds found in human speech patterns.^4,41,42 What the internal anatomy/microanatomy of their vocal folds was exactly like, and whether changes toward the human condition were occurring in the distribution, structure or innervation patterns of the intrinsic musculature, remains unknown. If the global reconstructions given above for their ADT are correct and they were indeed very similar to living apes, then it is reasonable to hypothesize that their vocal fold anatomy would also not be very different from those of the living apes and monkeys. They may well, for example, have lacked a deep layer of the lamina propria, again a feature suggested to be found only in living humans and not in other primates (Fig. 1.4).

If our earliest relatives were still “ape-like” in their ADT/vocal tract component, when did things change en route to us? This is a key question of our evolution and while definitive answers in evolution are rare, clues are beginning to appear. Our fossil data—and by extension reconstructions—suggest that the region was starting to change with the first members of our own genus, Homo, some million—plus years before the present on the plains of Africa.

While breathing and swallowing abilities were changing, so too were vocal and speech capabilities. On the one hand, the brain size in our early Homo ancestors was13 increasing over that found in earlier australopith-level hominids. Concomitant with that were indications that the internal complexity of the brain was changing as well. Together with the ADT changes noted above, the stage was set for vocal tract and intrinsic laryngeal change.

A watershed level of change was beginning with the advent of a new genus, one more advanced linguistically and cognitively than any that had preceded it. If, indeed, such changes were now becoming widespread, it is likely that the vocal component was taking on an ever-increasing communication role. Accordingly, microanatomical changes in vocal fold anatomy and function were also beginning to occur, as would have been alterations in intrinsic laryngeal muscle fiber type. While en route to the modern condition, our early Homo ancestors were probably not yet there. It is difficult, yet intriguing, to envision relatives who had an anatomical/physiological substrate different from either ourselves or our ape-like, earliest ancestors. The early members of our genus were remarkable in their own way, neither ape, nor australopith, nor us, humans, as we are today.

While changes may have begun with early members of our genus, it was not until the appearance of early members our own species, Homo sapiens, some 200,000 to 300,000 years before the present (or even much later depending on who one sees as worthy of H. sapiens designation) that our reconstructions indicate an ADT similar to ours appeared.⁴² These early H. sapiens, probably functioned much as we do today as regards their breathing, swallowing, and communicative parameters. Certainly, their extensive vocal tract component would have allowed for the production of the sounds of extant speech. We can also hypothesize that their internal laryngeal milieu was very similar to ours, but exactly when the type of muscle fibers, or presence of a lamina propria such as those shown today, actually appeared remains to be determined. Could there have been stages of change even within the timeframe of our own species? Could some of the answers to the late appearance of advances in the archeological record that point to highly advanced cognitive abilities⁴⁴ be reflective of subtle, yet seminal, changes in our voice or speech capabilities? Clearly, as we uncover more of the secrets buried within our unique structures for voice and speech, so too will be able to uncover their deeply entwined role with the story of how we came to be.