In Evolution in Four Dimensions, Jablonka & Lamb (J&L, Reference Jablonka and Lamb2005) provide a coherent, unifying evolutionary framework that does not compromise complexity but, instead, embraces it. From our perspective, the meatiest chapters of this book are those on genetic and epigenetic inheritance systems and their interaction. The authors' rich discussion of epigenetic inheritance is fascinating, and their criticism of pure selfish-gene theory compelling. Despite these positive impressions, we found J&L's discussion of behavioral and symbolic inheritance to be less thorough than their comprehensive coverage of genetics and epigenetics, from which we learned so much. Our primary concern is the lack of focus on the interaction between behavioral and symbolic inheritance. This limitation is surprising, given that gene–behavior and gene–language interactions are provided special attention. The relationship between genes and language may be more controversial than that between behavioral and symbolic inheritance. Nevertheless, we argue that behavioral–symbolic interactions are crucial for understanding symbolic language. A discussion of this interaction would have made J&L's four-dimensional framework more complete. In short, there is a chapter missing from this book. In what space we have, we propose some of the material such a chapter could contain.

First, inheritance through behavior-influencing substances may have an analogue in the prenatal auditory experience infants receive in the womb. Research by Shi and colleagues (Shi et al. Reference Shi, Werker and Morgan1999) suggests that very early auditory perceptual abilities may contribute to grammatical development. Shi et al. studied 1- to 3-days-old infants' responses to two separate classes of words: function words (short, unstressed words such as will and for) and content words, such as nouns and verbs. Infants detected a change when different words were presented auditorily, but only when the change happened across these classes. Other experiments have demonstrated perceptual sensitivity to the difference between the mother's voice and a stranger's voice in utero (Kisilevsky et al. Reference Kisilevsky, Hains, Lee, Xie, Huang, Zhang and Wang2003). J&L argue that maternal diet has chemical consequences that bias a child's early culinary sensitivities. Similarly, one may argue that maternal language has perceptual consequences that bias early linguistic sensitivities. Some may debate the functional significance of this phenomenon, but it is a reasonable possibility that it provides an advantage for children whose prenatal auditory experiences permit early preferences to the mother's language (DeCasper & Fifer Reference DeCasper and Fifer1980; DeCasper & Prescott Reference DeCasper and Prescott1984).

J&L's second behavioral inheritance system, non-imitative social learning, seems also to have a role in early language learning. Some developmental psychologists have argued that socially mediated environmental contingencies promote vocal and social learning that is crucial for later language skill. Watson's (Reference Watson1966; Reference Watson, Field and Fox1985) early work on infants' contingency perception offered a possible basis for emerging social skill. More recently, Goldstein and colleagues (Goldstein et al. Reference Goldstein, King and West2003) showed that, during early stages of vocal development, social contingencies can affect characteristics of vocalization, including the quality and quantity of vowel sounds. These contingencies likely emerge in the language-learning environment while the caregiver naturally attends to vocalization by the child (see also Goldstein & West Reference Goldstein and West1999).

Imitation, J&L's third type of behavioral inheritance, likely has some role in language development as well. In fact, numerous authors have argued for an intimate link between imitation and language (e.g., Meltzoff Reference Meltzoff1988; Tomasello Reference Tomasello2003). A great number of studies have suggested a variety of imitative behaviors by young children. Although debate continues about certain studies and their relevance, researchers have reported imitation of basic facial expressions and gestures early in infancy (Meltzoff & Moore Reference Meltzoff and Moore1977), sound productions that reflect the ambient language environment (de Boysson-Bardies & Vihman Reference de Boysson-Bardies and Vihman1991), and reproduction of novel words modeled by adults (Tomasello & Barton Reference Tomasello and Barton1994). These imitative actions may be foundational for aspects of language acquisition, including gesture, phonology, and word learning.

As children progress from early word learning to more advanced stages of language use, they seem to rely increasingly on symbolic inheritance to further their language skills. In formal education and informal learning alike, children and adults can learn the meanings of new words through multi-word definitions coming from direct instruction or even reference books. Thus, language knowledge is a product of both behavioral and symbolic inheritance (with the symbolic feeding back onto itself).

As J&L argue, there is a reasonable basis for distinguishing symbolic and behavioral inheritance. We urge, however, that this distinction can be subtle and deserves more attention, particularly when considering language and its acquisition. For example, is language learning transmitted mostly through symbolic inheritance, or through behavioral inheritance? At the early stages of learning, when the most fundamental linguistic conventions are developing, behavioral inheritance seems more relevant than symbolic inheritance. In fact, many have found it challenging to draw a cutoff where a child's language becomes symbolic. Language itself may be symbolic to varying degrees, depending on factors such as the extent to which a lexical item is generalized across environmental contexts, and the extent to which a phrase is conventionalized (Bates et al. Reference Bates, Benigni, Bretherton, Camaioni and Volterra1979; Bybee Reference Bybee2006; Tomasello Reference Tomasello2003).

We should note that, scattered throughout J&L's book, one can find reflections similar to those presented here. For instance, the authors discuss Chomskyan and functionalist theories about the role of input in language acquisition; they also suggest genetic adaptations in cognitive faculties that promote language learning. However, they rarely relate these reflections to their behavioral inheritance dimension. A formal discussion of behavioral–symbolic interaction would be useful for multiple reasons. First, behavioral inheritance underlies our ability to transmit symbolic information. Second, symbolic information presumably affects what is transmitted through behavioral inheritance. Finally, behavioral inheritance may be one route through which the genetic and symbolic dimensions interact.