Ackerman et al.'s plausible account of how brain evolution may have led to language would be even more persuasive if it also dealt with how evolutionary environments might have driven brain changes that engendered language. The survival and reproduction of organisms within populations depends on their behavior, so I start from the position that the brain, like all organ systems, evolved in the service of behavior (Catania Reference Catania2008). For example, brain size may have driven articulatory control, but environments where that articulatory control made a difference must also have driven brain size. Elsewhere I address in more detail these and related issues, including interpretations of learning in terms of selection rather than associations and the distinction between language structure and function (e.g., Catania Reference Catania1990; Reference Catania2013a; Catania & Cerutti Reference Catania, Cerutti, Thompson and Zeiler1986).
The functional distinction between affective language, as in tone of voice, and substantive language, as in vocal discourse, is illustrated by an account of the different reactions of two audiences to a speech by Ronald Reagan (Sacks Reference Sacks and Sacks1985). Psychotics without affect responded only to the speech content, whereas aphasics responded only to its affect; only those responsive to both dimensions found the speech persuasive. The affective and the discursive systems involve the same vocal apparatus, so they necessarily evolved in coordination, but Sacks's example demonstrates their separate functionalities. If affective vocal functions are similar to those of other displays usually characterized as emotional, then other functions must have driven evolution of the discursive system: evolution rarely duplicates existing functions. Emotional behavior substantially predates language, with discursive functions presumably overlaid upon it later. Contented or angry or lustful gorillas do not need new ways to express their emotions.
My candidate for the minimal function of discursive verbal behavior from which all other functions are derived? It is a highly efficient way in which one human can get another to do something (Catania Reference Catania1995; Reference Catania2003; Reference Catania2009). The imperative does not require multiple-word utterances or grammar. Even if nothing else is available, a single utterance functionally equivalent to the command Stop! will benefit the members of any hominid group that creates it. Other functions (e.g., communication, narrative) are derivatives of this fundamental one. For example, prestige matters only when some individuals become more important in telling others what to do; cooperation can sometimes be more effectively induced through verbal instructions than by other means. Telling others what to do leads to telling them what to say, and giving information provides expanded ways to tell others how to do things.
The long period during which our ancestors made tools and mastered fire suggests that some form of hominid language has existed for perhaps millions of years, whereas archaeological findings coupled with inferences from linguistic change and human migration implies a source perhaps as recent as 40 to 50 thousand years ago. The longer time makes sense if we include the evolution of affective and single-utterance precursors with simple imperative functions, accompanied by more sophisticated articulations. Significant anatomical developments included bipedal locomotion, freeing respiration from constraints on the rib cage, and elaborations of vocal signals such as laughter (Provine Reference Provine2000; Reference Provine2012). But defining language solely in terms of syntactically organized multi-word utterances targets the more recent provenance. The step from single- to multiword utterances with different words having different functions allows for an explosion of language diversity.
Coherent accounts of language evolution must include three concurrent levels of selection (Catania Reference Catania and Györi2001), each entailing different mechanisms by which environments select surviving variants. First, phylogenetic (Darwinian) contingencies must select requisite physiological attributes (e.g., vocal tract structure, neural organization). Second, ontogenetic contingencies (selection of behavior within individual lifetimes) must maintain those language features acquired by individuals, as when native but not non-native speech sounds survive in a child's developing vocalizations. Third, cultural or memetic selection (selection of behavior as it passes among individuals) must perpetuate languages across generations as communities pass them on from one to another.
Ackerman et al. seek an account for the co-evolution of articulatory and perceptual skills. Yet if distributions of both skill levels exist within a population, those at the upper ranges of either skill will be selected. As long as selection operates relative to each population mean, the distributions will change together, just as but more benignly than in the arms races of predators and their prey. For example, mothers with superior acuity along some auditory dimension will sometimes bear offspring with superior differentiation along some articulatory dimension; they and their offspring will both be selected, just as more successful predators are selected as their predation selects prey more successful at escape.
Ackerman et al. cite many relevant studies but provide no taxonomy of relevant processes. We read of reinforcement, goal-directed behavior, instrumental conditioning, learning responses to food rewards, acquisition of stimulus-driven behavioral routines, habit formation, training, and even motor tricks. But these are simply alternative vocabularies for labeling behavior changes that occur because behavior is modified by its consequences (Catania Reference Catania2013a; Schneider Reference Schneider2012). Consequences are as much involved in stimulus-driven behavioral routines, where responses produce different consequences given different stimuli, as in refinement of skills, where differential consequences shape behavior. Existing taxonomies of behavioral processes could put much of this in good order. They are built not upon associations or conditioning but rather upon the selection of behavior by its consequences (e.g., Catania Reference Catania2013a; Reference Catania2013b; Madden Reference Madden2012; Verplanck Reference Verplanck2000). The structure of behavior provides crucial clues for what to look for in the brain.
Nevertheless, Ackerman et al. have made a strong case that FOXP2's expression is prerequisite for the complex vocal articulations of human language. Timing is critical, so these coordinations involve not just tongue, lips, and larynx, but also diaphragm and rib cage. FOXP2's expression may work by allowing motor patterns that are otherwise highly constrained by anatomies and stimuli to be modified by their consequences. Skinner apparently got it right when he wrote: “The human species took a crucial step forward when its vocal musculature came under operant control in the production of speech sounds. Indeed, it is possible that all the distinctive achievements of the species can be traced to that one genetic change” (Skinner Reference Skinner1986, p. 117). In nonverbal organisms, reinforcers can alter only the rate of vocalizations (e.g., cheeps of chicks; Lane Reference Lane1961); in children, however, their form is sculpted even by such subtle differential consequences as producing sounds more or less resembling those of caregivers (Risley Reference Risley, Etzel, LeBlanc and Baer1977; Vihman Reference Vihman1996). This is as it should be because, as Ackerman et al. so effectively point out, our vocal articulations are perhaps the most sophisticated of human achievements.
Ackerman et al.'s plausible account of how brain evolution may have led to language would be even more persuasive if it also dealt with how evolutionary environments might have driven brain changes that engendered language. The survival and reproduction of organisms within populations depends on their behavior, so I start from the position that the brain, like all organ systems, evolved in the service of behavior (Catania Reference Catania2008). For example, brain size may have driven articulatory control, but environments where that articulatory control made a difference must also have driven brain size. Elsewhere I address in more detail these and related issues, including interpretations of learning in terms of selection rather than associations and the distinction between language structure and function (e.g., Catania Reference Catania1990; Reference Catania2013a; Catania & Cerutti Reference Catania, Cerutti, Thompson and Zeiler1986).
The functional distinction between affective language, as in tone of voice, and substantive language, as in vocal discourse, is illustrated by an account of the different reactions of two audiences to a speech by Ronald Reagan (Sacks Reference Sacks and Sacks1985). Psychotics without affect responded only to the speech content, whereas aphasics responded only to its affect; only those responsive to both dimensions found the speech persuasive. The affective and the discursive systems involve the same vocal apparatus, so they necessarily evolved in coordination, but Sacks's example demonstrates their separate functionalities. If affective vocal functions are similar to those of other displays usually characterized as emotional, then other functions must have driven evolution of the discursive system: evolution rarely duplicates existing functions. Emotional behavior substantially predates language, with discursive functions presumably overlaid upon it later. Contented or angry or lustful gorillas do not need new ways to express their emotions.
My candidate for the minimal function of discursive verbal behavior from which all other functions are derived? It is a highly efficient way in which one human can get another to do something (Catania Reference Catania1995; Reference Catania2003; Reference Catania2009). The imperative does not require multiple-word utterances or grammar. Even if nothing else is available, a single utterance functionally equivalent to the command Stop! will benefit the members of any hominid group that creates it. Other functions (e.g., communication, narrative) are derivatives of this fundamental one. For example, prestige matters only when some individuals become more important in telling others what to do; cooperation can sometimes be more effectively induced through verbal instructions than by other means. Telling others what to do leads to telling them what to say, and giving information provides expanded ways to tell others how to do things.
The long period during which our ancestors made tools and mastered fire suggests that some form of hominid language has existed for perhaps millions of years, whereas archaeological findings coupled with inferences from linguistic change and human migration implies a source perhaps as recent as 40 to 50 thousand years ago. The longer time makes sense if we include the evolution of affective and single-utterance precursors with simple imperative functions, accompanied by more sophisticated articulations. Significant anatomical developments included bipedal locomotion, freeing respiration from constraints on the rib cage, and elaborations of vocal signals such as laughter (Provine Reference Provine2000; Reference Provine2012). But defining language solely in terms of syntactically organized multi-word utterances targets the more recent provenance. The step from single- to multiword utterances with different words having different functions allows for an explosion of language diversity.
Coherent accounts of language evolution must include three concurrent levels of selection (Catania Reference Catania and Györi2001), each entailing different mechanisms by which environments select surviving variants. First, phylogenetic (Darwinian) contingencies must select requisite physiological attributes (e.g., vocal tract structure, neural organization). Second, ontogenetic contingencies (selection of behavior within individual lifetimes) must maintain those language features acquired by individuals, as when native but not non-native speech sounds survive in a child's developing vocalizations. Third, cultural or memetic selection (selection of behavior as it passes among individuals) must perpetuate languages across generations as communities pass them on from one to another.
Ackerman et al. seek an account for the co-evolution of articulatory and perceptual skills. Yet if distributions of both skill levels exist within a population, those at the upper ranges of either skill will be selected. As long as selection operates relative to each population mean, the distributions will change together, just as but more benignly than in the arms races of predators and their prey. For example, mothers with superior acuity along some auditory dimension will sometimes bear offspring with superior differentiation along some articulatory dimension; they and their offspring will both be selected, just as more successful predators are selected as their predation selects prey more successful at escape.
Ackerman et al. cite many relevant studies but provide no taxonomy of relevant processes. We read of reinforcement, goal-directed behavior, instrumental conditioning, learning responses to food rewards, acquisition of stimulus-driven behavioral routines, habit formation, training, and even motor tricks. But these are simply alternative vocabularies for labeling behavior changes that occur because behavior is modified by its consequences (Catania Reference Catania2013a; Schneider Reference Schneider2012). Consequences are as much involved in stimulus-driven behavioral routines, where responses produce different consequences given different stimuli, as in refinement of skills, where differential consequences shape behavior. Existing taxonomies of behavioral processes could put much of this in good order. They are built not upon associations or conditioning but rather upon the selection of behavior by its consequences (e.g., Catania Reference Catania2013a; Reference Catania2013b; Madden Reference Madden2012; Verplanck Reference Verplanck2000). The structure of behavior provides crucial clues for what to look for in the brain.
Nevertheless, Ackerman et al. have made a strong case that FOXP2's expression is prerequisite for the complex vocal articulations of human language. Timing is critical, so these coordinations involve not just tongue, lips, and larynx, but also diaphragm and rib cage. FOXP2's expression may work by allowing motor patterns that are otherwise highly constrained by anatomies and stimuli to be modified by their consequences. Skinner apparently got it right when he wrote: “The human species took a crucial step forward when its vocal musculature came under operant control in the production of speech sounds. Indeed, it is possible that all the distinctive achievements of the species can be traced to that one genetic change” (Skinner Reference Skinner1986, p. 117). In nonverbal organisms, reinforcers can alter only the rate of vocalizations (e.g., cheeps of chicks; Lane Reference Lane1961); in children, however, their form is sculpted even by such subtle differential consequences as producing sounds more or less resembling those of caregivers (Risley Reference Risley, Etzel, LeBlanc and Baer1977; Vihman Reference Vihman1996). This is as it should be because, as Ackerman et al. so effectively point out, our vocal articulations are perhaps the most sophisticated of human achievements.