In the Beginning Was the Familiar Voice Personally Familiar Voices in the Evolutionary and Contemporary Biology of Communication

Abstract

The human voice is described in dialogic linguistics as an embodiment of self in a social context, contributing to expression, perception and mutual exchange of self, consciousness, inner life, and personhood. While these approaches are subjective and arise from phenomenological perspectives, scientific facts about personal vocal identity, and its role in biological development, support these views. It is our purpose to review studies of the biology of personal vocal identity -- the familiar voice pattern-- as providing an empirical foundation for the view that the human voice is an embodiment of self in the social context. Recent developments in the biology and evolution of communication are concordant with these notions, revealing that familiar voice recognition (also known as vocal identity recognition or individual vocal recognition) or contributed to survival in the earliest vocalizing species. Contemporary ethology documents the crucial role of familiar voices across animal species in signaling and perceiving internal states and personal identities. Neuropsychological studies of voice reveal multimodal cerebral associations arising across brain structures involved in memory, emotion, attention, and arousal in vocal perception and production, such that the voice represents the whole person. Although its roots are in evolutionary biology, human competence for processing layered social and personal meanings in the voice, as well as personal identity in a large repertory of familiar voice patterns, has achieved an immense sophistication.

1. Voices Are All Around Us: Familiar Voices are Special

Voices are everywhere in human society. Every viable speaker in the world produces a signature voice pattern, and listeners naturally derive an abundance of judgments from it. All kinds of consequences, benign as well as dire, flow from human interactions with voice quality. Understanding these judgments and analyzing these behaviors have interested researchers in an appropriately large array of disciplines, including psychology, linguistics, speech and hearing sciences, forensics, communicative disorders, evolutionary biology, and anthropology. This is true because voices exist and operate in personal, sociological, and historical dimensions. Although these dimensions are manifest only in the speaker-listener interaction, in traditional approaches to the study of voice, speakers and listeners are examined as if they were independent entities linked only by a physical (acoustic) signal. Yet the voice pattern is best approached as constituting a dialogic process. As such, it has evolved to play a major part in the biology of communication, signaling reproductive fitness, fostering mother/infant reunifications and bonding (and hence infant survival), facilitating identification of friend and foe, and enabling the formation of social groups (Locke, 2008, 2009; MacLean, 1990). Studies of this theme often utilize the term “vocal recognition of identity” (e.g., Rendall, Rodman, & Emond, 1996). Here the term “familiar voice pattern” is intended to emphasize the roles and properties of this important feature of biology. It is our purpose to describe how it is the familiar voice—known and recognizable to the listener—that has most prominently played these crucial roles in the biology of communication.

This article provides biological foundations to the characterization of voice as embodied social form representing a person and tied to the mutual sharing of self and consciousness (e.g., Bertau, 2007, 2008; Osatuke et al., 2004; Linell, 2007; Hermans, 1996, 1998; Shotter, 1996). Related to this view is the elaboration of the multilayered, polyphonic nature of the personal voice pattern (Stiles, 1999; Stiles, Osatuke, Glick & Mackay, 2004) which has its finest flowering, in interspeaker dialogue, when the familiarity aspect is present. These conceptualizations of voice quality--physical embodiment and multilayering as revealed in the dialogic process of mutually sharing, speaking and listening subjects---are supported by evolutionary and biological perspectives (Kreiman & Sidtis, 2011). Evidence is provided here, first, that familiar voice patterns are special in human affairs; that their salient role in infant survival begins even before birth; that inherent in each is an elaborate constellation of biographical information; and that it takes the whole brain and, by extension, the whole person to participate in producing and perceiving a voice. Biological perspectives reveal that synergistic production and perception of familiar voices in the environment have crucially guided survival and development across the millennia of evolution. These mutually shared vocal exchanges arise from the specific cultural contingencies of social groupings (Bakhtin, 1973, 1986; Neisser & Jopling, 1997; Josephs, 2002).

Voice quality processing in humans is a prodigious cognitive ability, second only to language in depth, complexity, and extent. Processing of the voice pattern evolved over millennia of development in many species, but the elaboration of the role of voice in humans is immense, just as human language has also attained a certain unique complexity. Our research leads to the proposal that it is the personally familiar voice that emerges as crucial to biological, evolutionary and social development. No known limit has been demonstrated for the repertory of recognizable voices in humans. Acquaintances expect to be recognized by voice (Schegloff, 1979). It is likely that other vocal information—intentions, emotions, inferences, mood, attitudes—is more competently transmitted when familiar voices are in play (Nygaard, 2005). Production and perception of personally familiar voice patterns, representing the multifaceted self in significant behavioral and cultural contexts, antedated by a very long time the development of speech and language, and allowed for social relating between individuals in earliest evolutionary times. Better understanding of the role of familiarity in human neuropsychology has come only recently, because utilizing familiar stimuli in controlled studies poses special challenges (Van Lancker, 1991; Kreiman & Sidtis, 2011).

One might suppose that recognizing a familiar voice arises from a more fundamental ability to discriminate among unfamiliar voices (e.g., Fischer, 2004). However, convergence of several observations suggests that familiar voice recognition is the more elemental process. First, persons with focal brain damage are able to recognize familiar voices despite a neuropsychological impairment in the ability to discriminate among unfamiliar voices (Neuner & Schweinberger, 2000; Van Lancker & Kreiman, 1986, 1987; Van Lancker, Cummings, Kreiman, & Dobkin, 1988). The reverse is also true, suggesting that in the adult, the two abilities may exist as independent and unordered in processing, and that, although both abilities are based in fundamental features of auditory-acoustic processing, recognition does not necessarily depend on or arise from discrimination. Interactions of the two functions at several stages of auditory processing and during acquisition and development, of course, are to be expected.

Secondly, as further evidence of the primacy of the familiar voice, the ability to recognize the voice of one's mother is present at birth in normally hearing humans (DeCasper & Fifer, 1980; Querleu et al., 1984; Hepper et al., 1993; Mehler, Bertoncini, Barriere, & Jassik-Gerschenfeld, 1978), in contrast to the ability to discriminate among unfamiliar voices, which does not fully develop until about age 10 (Mann, Diamond, & Carey, 1979). The role of mother-infant mutually instantiated vocal recognition in establishing personal ingredients as building blocks of self, social development and acculturation can hardly be exaggerated (Bertau, 2007). Neurophysiological evidence confirms that the mother's voice has special meaning to an infant, in that mother's voice excites unique brain patterns (Purhonen, Kilpelainen-Lees, Valkonen-Korhonen, Karhu, & Lehtonen, 2004, 2005). Recognizing and responding to the mother's voice is an important ingredient for fostering infant/mother bonding, which in turn is essential for the survival of the infant given its total dependence on its mother (Locke & Bogin, 2006). In fact, recognition of mothers by infants, infants by mothers, and mutual voice recognition between parents and offspring are all common in animals, as discussed below (Section 3).

Beyond early infancy, familiar voices retain their special status as signifiers of multilayered information. The personally familiar voice as an emotional auditory event travels via specialized auditory pathways directly to subcortical mechanisms, which, according to the Cannon-Bard theory, arouse corporeal and cerebral representations of emotion (Damasio, 1994). The voice carries and evokes continuous flickers of emotion and attitude, which arise from associated evaluations stored in memory that “become active automatically on the mere presence or mention of the object in the environment” (Bargh, Chaiken, Govender, & Pratt, 1992, p. 893). Implicit memorial associations of dizzying variety attach to the familiar voice pattern (Figure 1). In contrast, inner experiencing of an unfamiliar voice is relatively impoverished. The subjective emotions associated with familiar voices, together with the prosodic qualities carried in the voice signal at the time of hearing the voice, combine to form a powerful vehicle to acquire, process, and recall the personally familiar voice pattern (Laird, Wagener, Halal, & Szegda, 1982). (We have elsewhere described the rapid acquisition of the familiar voice pattern, due to engagement of arousal and emotional systems in the brain (Kreiman & Sidtis, 2011, Chapter 6).) These ingredients contribute to the expression of self and personhood in the voice and to the mutual sharing of self in communicative interaction.

Figure 1.

A subjective, unstructured array of qualities accompanies the known voice pattern as personally relevant factors.

2. It Takes a Whole Brain to Produce and Recognize a Voice

The whole self underlies the personal vocal pattern, such that a large range of subtle, ephemeral personal characteristics are encoded and conveyed to the listener, who infers these attributes from voice. This includes the physical self (gender and age, health, reproductive fitness, race, size, and attractiveness; see Kreiman & Sidtis, 2011, chapter 4), the social self (education, background), and the speaker's personhood (personality, mood, emotions, and attitudes) (Laver, 1968; Berry, 1990; Revelle & Scherer, 2009; Scherer, 1986; Konopczynski, 2010). Producing and perceiving these characteristics relies on disparate brain structures that modulate cross-modal associations, memory, attention, arousal and emotion (Andersen, 1997; Fair, 1988, 1992; Kayser & Logothetis, 2007; Schroeder, et al., 2003). These facts provide the foundation for the notion that the voice is revelatory of “self,” mental states, and consciousness, and reflects both the speaker and the context in which the voice is produced. The association of voice with self, consciousness, and in earlier times, with soul, enjoys a venerable tradition since at least the time of Aristotle (von Kempelen, 1791; Rush, 1823; Kidd, 1857; Konopczynski, 2010) and is enfolded in the goals of contemporary voice coaching and therapy (e.g., Linklater, 1976; Boone, 1991). Modern perspectives describe how vocal signals reflect the listener's assumptions, expectations, experiences with the speaker, and cognitive perspectives (Pollermann, 2010). In this conceptualization, it is largely through use of the voice that the self of each co-participant is mutually shared in communicative interaction (Berteau, 2008); the embodied voice arises from the whole person. All dimensions of the voice enter into this process for the listener, but little neuroscientific research has focused on the manner in which voice arises from such complex social and psychological processes, so that much of the communication that passes between interlocutors via the voice remains undescribed and unexplained. In this section, we provide grounding for these ideas derived from recent developments in cerebral function.

In describing the process of vocal production, studies of voice usually focus on a few cerebral structures: the motor strip of the cortex, selected motor pathways through the basal ganglia, and a series of cranial nerves that activate laryngeal and vocal tract muscles. It is more accurate to say that producing a voice in humans involves, in addition to these structures, the midbrain, most subcortical nuclei and limbic structures, the cortical lobes and the cerebellum (Kreiman & Sidtis, 2011; Jürgens, 2002). One source of this proposition arises directly from the effects on voice of neurological damage or experimental stimulation: the periaqueductal gray matter in the midbrain (Esposito, Demeurisse, Alberti, & Fabbro, 1999), several basal ganglia and limbic (subcortical) structures (Simonyan & Jürgens, 2003; Damasio, 1994), temporal, parietal and frontal lobes, and the cerebellum are all involved in normal voice production (Ackermann, Mathiak, & Riecker, 2007). Indirect evidence lies in the respective neuropsychological roles of these structures: initiating behaviors, modulating motor gestures and emotion, monitoring auditory input, engaging pertinent cognitive associations (Fair, 1992), and managing ongoing action (Masterman & Cummings, 1997). Similarly for voice perception, traditional neuroanatomical descriptions depict the auditory pathway as a through-put channel and the auditory receiving areas of the cortex as providing interpretation of sound. In fact, perceiving a voice engages many levels of processing all along the auditory pathway, and visual, somatosensory and auditory signals play key roles at early, low-level stages of auditory cortical processing (Schroeder et al., 2003). This confluence of auditory and multisensory streams precedes cognitive processing of sound (Winer & Lee, 2007). It is now recognized that sound processing in the auditory cortex occurs not in one receiving center, but in various cortical fields, and further, that these areas interact richly with other areas (Petkof et al., 2006; Winer & Lee, 2007) through numerous connections with subcortical nuclei and cross model association areas of the frontal and parietal lobes (Fair, 1992; Benson, 1994; Haramati, Soroker, Dudai, & Levy, 2008). This picture of interactional, whole brain processing of sound is compatible with current models of the voice as infused with affect, thought, memory, motivation and attention, via a web of reciprocal influences involving multisensory areas and nuclei along with extensive cerebral connections (Miall, 1986). Thus the voice pattern arises from coordinated systems in the whole brain, which manifest the collective perspectives, knowledge domains and experiences of the speaker, deriving from an integrated network of limbic, cortical, subcortical, cerebellar and brainstem structures that underlie and modulate vocal production and perception and contribute to each produced or perceived voice pattern.

While neuropsychological research was focused for several decades on the contributions of the cortex to human behavior, there is currently a resurgence of interest in limbic structures, the seat of emotion, with neuroscientists pointing out that all of human behavior is richly imbued with emotional tone (e.g., Panksepp, 1998, 2003). Interest in subcortical structures, which are now known to modulate motivation, attention, and initiation of motor behaviors, has also grown (Marsden, 1982; Bhatia & Marsden, 1994; Saint Cyr, Taylor, & Nicholson, 1995; Lieberman, 2002). This perspective is concordant with current descriptions in neuroscience of “large-scale neurocognitive networks” contributing to mental activity that is highly flexible and almost infinitely rich (Mesulam, 1990, p. 597), and with a movement away from localization of functional representation to descriptions of functional networks in the brain (Nudo, Plautz, & Frost, 2001). This modern representation of voice production and perception is consonant with dialogic conceptualizations of voice as reflective of self in the broadest sense and with multilayering of personal information (Osatuke, Humphreys, Glick, Graff-Reed, McKensie Mack, & Stiles, 2005; Bertau, 2007a,b; Hermans, 1998).

3. In the Beginning Was the Voice Pattern: Voice Recognition Evolved Early and Across Many Species

Ethological and evolutionary data indicate that, in addition to use of various call repertories for semantic signaling purposes, recognition of familiar voice patterns, which include crucial information about the physical size, gender, emotions, mood, and subtleties of intention, is a widespread ability that has evolved in response to different environmental and behavioral demands. Monkeys and primates, bats, penguins, sheep, goats, deer, horses, wolves, frogs, elephants, and birds recognize the familiar voices of parent and/or child, or of a neighbor. Based on its modern distribution, this ability had evolved by the time that frogs appeared and before the advent of mammals (Burke & Murphy, 2007; Bee, Kozich, Blackwell, & Gerhardt, 2001). The males of a variety of anuran species (including the North American bullfrog, aromabatid frogs, and concave-eared torrent frogs; Bee & Gerhardt, 2002; Gasser et al., 2009; Feng et al., 2009) are able to discriminate the voices of familiar neighbors from those of unfamiliar frogs, and use this information to determine the level of aggressive response necessary to defend their mating grounds. Males respond significantly less aggressively to a familiar voice than to an unfamiliar voice, independent of the location from which the call sounds, suggesting that they can discriminate the familiar voices of their neighbors from unfamiliar voices, and can use this information to limit unnecessary aggressive interactions (Bee & Gerhardt, 2002).

Producing and recognizing familiar voice patterns thus antedates by millions of years the other more celebrated evolutionary developments in communication and cognition. Although voice recognition abilities are shared by many animals, the manner in which animals recognize one another varies widely in form and complexity from species to species. In recent times, vocal identity studies, or recognition of the individual by voice, have proliferated in the scientific literature (see Kreiman & Sidtis, 2011). In some species, voice is the primary means by which individuals recognize each other, as in penguins (Jouventin, 1982); or it may be used as an adjunct to visual and/or olfactory cues, as in sheep (Searby & Jouventin, 2003) and goats (Terrazas, Serafin, Hernandez, Nowak, & Poindron, 2003). Recognition may be mutual, so that infants and parents recognize each other, or asymmetrical, so that parents recognize offspring or offspring parents, but not both. For example, fallow deer fawns hide from predators, and thus must remain silent until summoned by their mothers; thus, fawns can recognize the calls of their own mothers but mothers do not recognize fawns (Torriani, Vannoni, & McElligott, 2006). In contrast, ewes and their lambs, who live in herds, can each recognize the other based solely on their calls (Searby & Jouventin, 2003). Finally, calls can be relatively simple in structure in animals that breed in small herds, follow their parents (as sheep and goats do), or rely on additional cues to recognize family members. For example, the exact role that voice plays in recognition varies across penguin species, and call complexity varies accordingly (Jouventin & Aubin, 2002; Searby, Jouventin, & Aubin, 2004). In nesting (Adélie, macaroni, and gentoo) penguins, the nest location is the primary cue to the penguin's identity, so calls serve mostly to summon a chick to the nest when a parent returns from foraging and are accordingly quite simple in acoustic structure, with recognition depending primarily on pitch (Jouventin & Aubin, 2002). In contrast, emperor and king penguins incubate their eggs on their feet, so recognition depends completely on voice quality. Accordingly, their calls are acoustically highly complex (Aubin, Jouventin, & Hildebrand, 2000; Lengagne, Lauga, & Aubin, 2001).

Various types of voice patterns and recognition strategies have evolved in other species that breed in colonies, produce mobile offspring, and/or forage, and thus must rely on voice recognition if offspring are to survive (e.g., Searby et al., 2004). For example, seals breed in colonies of up to 70,000 animals, and because feeding grounds are often far from breeding shores, mothers must leave pups unattended sometimes for weeks while foraging for food. Mothers recognize pups by voice in seven different species of seal; in four of these, pups also recognize their mothers (Insley, 2001). While pups' voices change as they grow, mother seals retain the ability to identify these different versions produced at different ages of development (Charrier, Mathevon, & Jouventin, 2001, 2003). Voice patterns may also be inborn: Evening bat mothers go out to forage immediately after the birth of their young, but successfully reunite with their pups in dark crèches that may contain thousands of other bats, suggesting calls have a genetic component (Scherrer & Wilkinson, 1993). Although unilateral or mutual voice recognition is usually instantaneous for offspring, learning to recognize a parent's or infant's voice may follow a delayed course in animals whose environment and/or behavior makes voice recognition less critically important. In some primate species the responsibility for reunification rests on mothers, and infants may not recognize their parent until well after birth (Altmann, 1980). Japanese macaques do not recognize their mothers until about 22 days of age (Masataka, 1985), and infant barbary macaques apparently do not recognize their mothers reliably until 10 weeks of age (Fischer, 2004). Although studies in the field confirm familiar voice pattern recognition behaviors in nonhuman primates (Cheney & Seyfarth, 1980, 1999; Rendall, Rodmann & Edmond, 1996), the repertory of recognized identities in other species studied up to this time is usually limited to a very few at a time (Hansen, 1976). One exception lies in elephant society, where many individuals in surrounding herds—up to a hundred-- are recognized by voice (McComb, Moss, Sayialel, & Baker, 2002). Further studies may reveal more about this ability in nonhuman animals. In the meantime, human management of exquisitely layered prosodic meanings within an immense expanse of personal identities, signaled by the voice pattern, remains an astonishing and dazzling biological tour de force.

In summary, the wide-spread presence and sophistication of familiar voice recognition across species, along with the variety of recognition behaviors that has evolved, attest to the fundamental importance of this ability in biology (MacLean, 1990). The details of the observed variations can often be understood in terms of the varying ecological and evolutionary demands faced by different animal species. These facts underscore the importance of the ability to recognize a familiar voice, the essential mutuality of the process, and the special status that familiar voices have in biology and animal behavior.

4. A Neuropsychological-Social Model of Voice Recognition

Acquiring and recognizing a familiar voice involves tapping into an extensive network of attributes associated with the target voice, and draws on a constellation of diverse characteristics, including affective and attitudinal qualities, biographical history, appearance, dress, gait, unique markings and customary paraphernalia, preferences, and so on, which are stored together in a broad based, integral Gestalt percept (see Figure 1). Subjective qualities in philosophical treatments of consciousness, self, and awareness (Buck, 1993) emerge most prominently for the familiar voice pattern, as essentially different from the status and role of unfamiliar voices. A final common pathway for vocal pattern recognition occurs in the right hemisphere, site of personal relevance (Van Lancker & Canter, 1982; Van Lancker & Kreiman, 1986; Belin, Zatorre, Lafaille, Ahad, & Pike, 2000; von Kriegstein & Giraud, 2004; Sidtis & Kreiman, 2009). In contrast, discriminating unfamiliar voices utilizes feature-analysis and feature-matching to a general template or set of templates that approximately fit a set of perceived auditory features, and is more successfully modulated by the left cerebral hemisphere and/or by both hemispheres (Kreiman, 1997; Kreiman & Sidtis, 2011). This characterization of voice perception is concordant with known differences in cerebral processing emerging from clinical and experimental data, with familiarity and pattern recognition represented in the right hemisphere, and temporal and analytic processes more successfully processed in the left cerebral hemisphere (Bever, 1975; Bradshaw & Mattingly, 1995; Gazzaniga, Ivry & Mangun, 2002; Van Lancker, 1997).

Loosely adapting psychological approaches that contrast exemplar and rule-based models of learning and memory, we describe these differences with a “Fox and Hedgehog” model of voice recognition, named after the parable in which the fox knows many little things, while the hedgehog knows one big thing (Berlin, 1953/1994). This model proposes that perceptual-auditory features (lots of little things) are utilized predominantly for perceiving unfamiliar voices, while for recognition of a familiar voice, a few signature features suffice to herald the complete, known voice pattern (one big thing) (Kreiman & Sidtis, 2011, Chapter 6).

This view of vocal quality perception is consistent with microgenetic or process theory (Brown, 1998a, b), which provides a distinctive counterpart to paradigms of brain function based on assumptions about localization of neuropsychological abilities. Microgenesis refers to psychodynamic processes unfolding in a present-time scale from global to local instantiation of a realized percept or experienced mental event (Werner, 1956; Rosenthal, 2004). A key feature is that form, meaning and value are not independent, but unfold simultaneously from resources within the entire brain (Brown, 1988). Our perspective on familiar voices is resonant with the process model of brain function in which configurations play a major role as original status of the cognitive content (Benowitz, Finkelstein, Levine, & Moya, 1990; Brown, 1998b). The mental content is “not constructed like a building,” but unfolds from “preliminary configurations [that] are implicit in the final object” (Brown, 2002, p. 8). Affect and familiarity, essential characteristics of familiar voice patterns, are better accommodated in these approaches (Brown, 1998b, 2002), which lend themselves to a discussion of affective, subjective, and personally familiar phenomena (Van Lancker, 1991). This approach provides a vehicle for characterizing the intimate listener-speaker dyad inherent in familiar voice processing, and for the fact that the voice is expressive of the entire self.

5. Familiar Voice Patterns Enable Relationships in Sentient¹ Life

Voice recognition competence, or successful recognition of vocal identity, is present in many species and present at birth in humans, and as an evolutionarily “old” competence takes its place among the most primordial of behaviors. Following Gibson's (1966) view that sensitivity to biologically useful information evolves with the organism, it follows that sensitivities for acquiring a familiar voice pattern arise from the most basic perceptual processes.

The pervasiveness and importance of the human voice in the psychology and sociology of life, and the breadth of its effects on human behavior, cannot be overstated, and the special status of the familiar voice pattern has been described here². There are seldom events in daily living that do not involve production and perception of voice, with the attendant conscious and unconscious judgments of the talker by the listener. Personally relevant voices, by definition, are represented in memory with emotional reference to the self. Subjective impressions of voice as embodiment of self are strongly supported by the biological facts presented here. The intense personal meanings of voice have a profoundly important role in animal behavior, and continue to weave subtle strands of communication everywhere in modern human life. Social relating across biological species has relied on familiar voice patterns since the very beginnings of vocal behavior. Friend or foe, offspring and parents, family and cohort members can be identified at night or at a distance not by visual inspection, but by the voice pattern. The meaning and use of the vocal pattern grew with the growth in the complexity of cognition and social organization of various species. The prodigious presence of familiar voice patterns in human experience, and its role in facilitating mutual exchange of the inner voice, has its foundations in an evolutionary past shared by many other animals. However, the evolutionary leap from observable animal behaviors to the present state of human competence in voice recognition is as great as that from animal calls to human language. In humans, as described by Bertau (2008), the evolution of the personally relevant voice pattern accompanies and enhances the development of consciousness and self-awareness, as well as empathy for and recognition of the other.