Rachlin overlooks that free will determines when and in what direction acts that appear impulsive will occur. Because behavioral patterns continuously evolve, animals are not guaranteed when they will, or how to, maximize larger-later reinforcements. An animal therefore uses self-control to emit free acts to vary behavioral patterns to optimize larger-later rewards.

The costliness of disrupting a pattern may not be a useful criterion for distinguishing between acts and patterns; there are instances in which omitted components of patterns are hard to detect (e.g., typographical errors), or in which distortions are easily introduced (e.g., slurred words in a trite phrase). Are there behavioral criteria for distinguishing between acts and patterns?

The Aristotelian position on causality advocates explaining biological events by the configuration of outcomes they normally achieve. Rachlin's analysis follows this strategy as an alternative to explanation via hypothetical “inner” events, though he omits considering one definition of “inner” explanation (historical definition) as being complementary to teleological explanation. He also fails to provide a detailed teleological analysis of the role of bridging responses in allowing animals to select larger-outcome alternatives.

Operant psychologists are looking for alternatives to radical behaviorism. Rachlin offers teleological behaviorism, but it may pose as many difficulties as radical behaviorism. There is, however, a less drastic way to defend Rachlin's thesis of “order in the relationship between organisms and environment.” It portrays operant principles as relating distal efficient causes to behavioral effects.

The willingness to trade off large but ill-defined future consequences for immediate work characterizes social problems such as environmental sustainability. This commentary argues that important applications of behavioral models of self-control are being overlooked in the experimental literature. Tying the experimental literature to longterm health, environmental, and other risks makes the experimental work more germane, and raises new research questions for experimental modeling.

The target article extends molar behaviorism in two positive ways: beyond average aggregates and beyond restricted laws of “conditioning.” Although a molar framework based on purely overt events shows promise for advancing behavior theory, Rachlin's specific form of teleological behaviorism is in need of clarification.

Rachlin's teleological behaviorist account shows promise for extending our behavioral understanding of many aspects of self-control, particularly when self-control involves controlling behavior through larger behavioral patterns. However, it may face special challenges in dealing with instances of self-control that involve pattern breaking.

Rachlin's view of self-control is rejected on the grounds that his arguments do not establish the possibility of “molar” abstract, external, stimulus patterns and that his experiments, although they show that pigeons and human beings do sometimes choose postponed rather than immediate gratification, do not challenge the commonly held view that internal factors are involved in the former choice.

Rachlin proposes that the mechanism underlying self-control involves the cost incurred in changing one's response pattern, but his account explains neither how the response pattern is originally established nor why a change in response pattern entails a cost. I suggest that response patterns occur because they tend to be effective (thus alternatives are more costly) and are maintained because the decision process itself is costly.

Complex cases of self-control involve processes such as guilt-avoidance, inhibition, self-punishment, conscious thought, free will, and imagination. Such processes, conceived as internal mediating mechanisms, serve the function in psychological theory of avoiding teleological causation. Acceptance of the scientific legitimacy of teleological behaviorism would obviate the need for internal mediation, redefine the above processes in terms of temporally extended patterns of overt behavior, and clarify their relation to selfcontrol.

This target article presents a plausible evolutionary scenario for the emergence of the neural preconditions for language in the hominid lineage. In pleistocene primate lineages there was a paired evolutionary expansion of frontal and parietal neocortex (through certain well-documented adaptive changes associated with manipulative behaviors) resulting, in ancestral hominids, in an incipient Broca's region and in a configurationally unique junction of the parietal, occipital, and temporal lobes of the brain (the POT). On our view, the development of the POT in our ancestors resulted in the neuroanatomical substrate consistent with the ability for representations in modality-neutral association cortex and, as a result of structure-imposing interaction with Broca's area, the hierarchically structured “conceptual structure.” Evidence from paleoneurology and comparative primate neuroanatomy is used to argue that Homo habilis (2.5–2 million years ago) was the first hominid to have the appropriate gross neuroanatomical configuration to support conceptual structure. We thus suggest that the neural preconditions for language are met in H. habilis. Finally, we advocate a theory of language acquisition that uses conceptual structure as input to the learning procedures, thus bridging the gap between it and language.

Despite some sound basic assumptions, Wilkins & Wakefield portray a Homo habilis too linguistically sophisticated to fit in with the subsequent fossil record and thereby lose a reasoned explanation for human innovativeness. They err, too, in accepting a single-level model of conceptual structure and in deriving initial linguistic units from calls, a process far more dubious than the derivation of home-sign from naive gesture.

Preadaptation for language is an unnecessary assumption because intermediate stages of linguistic ability are possible and adaptive. Language could have evolved through gradual selection from structures exhibiting few features associated with modern structures. Without physical evidence pertaining to language ability in prehabilis hominids, it remains possible that selective pressures for language use preceded and necessitated modern neurolinguistic structures.

Wilkins & Wakefield assign importance to motor systems but skip from anatomy to cognitive structure with little attention to behavior. Organisms, no matter how sophisticated, that do not behave in accord with what they know will fall by the evolutionary wayside. Facts about behavior can supplement the authors' theory, whose hierarchical structures can accommodate an evolutionary scenario in which a million years or more of functionally varied utterances mainly limited to single words is followed by an explosion of linguistic diversity with the development in the last 50,000 years or so of syntactically organized multiple word utterances.

The precise manner in which language serves its communicative function suggests that natural selection, rather than exaptation or reappropriation, played the major role in its evolution. Natural selection is more readily invoked, I suggest, if it is assumed that language originated as a system of manual gestures, and later switched to an oral mode.

Three issues are addressed in this commentary. (1) Wilkins & Wakefield are commended for placing the complex behavior they discuss within an evolutionary matrix. (2) They err on a number of points in regard to their treatment of this complex behavior. These involve (a) their emphasis on the evolution of conceptual structure rather than language, (b) their equation of meaning with reference, (c) their minimalist view of learning theory, and (d) their separation of the evolution of speech from that of language. (3) They adopt a framework for neural preconditions that has been clearly discredited on both behavioral and neuroanatomical grounds. Finally, recent research involving modern neuroimaging techniques casts further doubt on the authors' postulated functions for the neural areas they discuss.

Hominid-like morphology in habiline cranial endocasts does not necessarily imply the presence of language capacity. The cortical zone in question is not associated exclusively with language in humans, and its emergence in habilines might indicate the evolution of other cognitive functions special to humans that were preconditions for the later evolution of language.

Wilkins & Wakefield suggest that changes in the hominid brain made it uniquely “preadaptive” for language, yet no precursor functions served as adaptive substrates to the emergence of language. We present contrary evidence that the ability to discriminate and process rapid and complex auditory information is a cross-species function subserving communication processes including, but not limited to, human speech perception. We suggest that auditory temporal processing served as an evolutionary precursor to speech processing and consequent language development in humans.

Wilkins & Wakefield fall short of solving the language origin puzzle because they underestimate the cognitive and linguistic capacities of great apes. A focus on ape capacities leads to the recognition of varied levels of cognition and language and to a gradualistic model of language emergence in which early hominid language skills exceed those of the apes but fall far short of those of modern humans or later fossil hominid groups.

Contra Wilkins & Wakefield, we argue that an evolutionarily inspired approach to language must consider different facets of language (i.e., more than syntax and semantics), and must explore the possibility of nonhuman precursors. Several examples are discussed, illustrating the power of the comparative approach in illuminating our understanding of language evolution.