Christiansen & Chater (C&C) are correct in their claim that it is implausible that language as we know it is the product of biological evolution for a complex, language-specific endowment. As they point out, this is not in itself an unusual claim. For example, Hauser et al. (Reference Hauser, Chomsky and Fitch2002) argue that the sole language-specific capacity might be recursion, and that even recursion might not be language-specific. However, our reasoning to this claim is quite different from C&C's. Language as we know it requires advanced capacities for meaning construction. Any view of language as having been active in quasi-advanced forms before about 50 thousand years ago, advancing further by refinements, runs up against the unlikelihood that human beings who enjoyed these near-modern capacities of meaning construction through language (blending, hypothetical and counterfactual thought, negatives, framing, mappings of many sorts, viewpoint and perspective, metaphor, speech acts, deixis, etc.) failed to use them to achieve even the rudiments of the other human higher-order cognitive achievements that we know require those very same mental resources: art, music, mathematical and scientific discovery, religious practices, representation, fashion, advanced social cognition, and so on.

The archeological record (Klein Reference Klein1999; Mithen Reference Mithen1996), genetic record (Cavalli-Sforza Reference Cavalli-Sforza2000; Thompson et al. Reference Thomson, Pritchard, Shen, Oefner and Feldman2000), and modern understanding of the uniformity of conceptual mappings in multiple, superficially diverse areas of human behavior all converge to indicate that biological evolution slowly produced not language but, rather, capacities not specific to language, evolving along a cline of increasing power. Once those capacities reached a certain sine qua non stage on the gradient, language and other human singularities became possible (Fauconnier & Turner Reference Fauconnier and Turner2002; Reference Fauconnier and Turner2008; Turner Reference Turner1996).

One (and perhaps the) mental operation whose evolution was crucial for language is conceptual blending (Coulson Reference Coulson2006; Coulson & Oakley Reference Coulson and Oakley2000; Reference Coulson and Oakley2005; Fauconnier & Turner Reference Fauconnier and Turner1998; Reference Fauconnier and Turner2002; Turner Reference Turner1996). Conceptual blending is a fundamental mental operation whose rudimentary forms are evident in the mammalian line of descent. The most advanced form, double-scope blending, consists of integrating two or more conceptual arrays as inputs whose frame structures typically conflict in radical ways on vital conceptual relations, such as cause-effect, modality, participant structure, role-value, and so on, into a novel conceptual array whose frame structure draws selectively from the frame structures of the inputs and dynamically develops emergent structure not found in either of the inputs.

There is now massive evidence that double-scope blending is indeed a necessary feature of all the human singularities mentioned above, including language. Conceptual blending itself has been evolving since at least deep in the mammalian line. Once the stage of double-scope blending was achieved, the full range of human higher-order singularities became possible. They arose in concert, reinforcingly, in cultural time. The archeological record suggests that the stage of double-scope blending was reached not millions of years ago and not even at the stage of anatomical speciation for human beings (about 150 thousand years ago), but rather, more recently, perhaps 50 to 80 thousand years ago (Mithen Reference Mithen1996).

Grammatical constructions are products of conceptual mapping. Specifically, they are double-scope blends (Fauconnier & Turner Reference Fauconnier, Turner and Goldberg1996; Liddell Reference Liddell2003; Mandelblit Reference Mandelblit1997). Accordingly, the existence of language requires the capacity for double-scope blending. Because double-scope blending is necessary for these patterns of meaning construction, because grammatical constructions in general are not possible absent the capacity for double-scope blending, and because double-scope blending is necessary for equipotentiality – the amazing ability of language to be used effectively in any situation, not just those that fit a finite list of frames (Fauconnier & Turner Reference Fauconnier and Turner2002) – language (like other human higher-order conceptual singularities) appears only once mapping capacities have evolved to reach the stage of double-scope blending. Intermediate stages of the mapping capacities are useful and adaptive, but not for language, which demands equipotentiality. This explains the absence of intermediate stages of language as an observable product.

There is no evidence known to science of simple languages now or at any time during our phylogenetic descent, because there are no simple or rudimentary languages. Once the stage of double-scope blending is achieved, fully complex language comes on like a flood, in cultural time rather than in biological time. In the early twentieth century, scientists at last gave up the argument that the cave paintings of France and Spain could not date from the Upper Paleolithic. They had assumed that art must have gone through stages of increasing complexity and accomplishment, that it must have been simple at its birth. We argue that a century later, scientists must give up the same argument and the same mistaken assumptions about language.

There is evidence of simpler stages, not of language or art, but of the mental operation underlying these behaviors – conceptual integration. The more rudimentary forms of conceptual integration remain available to all human beings, who moreover use those simpler forms constantly. Arguably, the evolution of mapping capacities in human beings was adaptive, and took place in standard fashion, but left no “fossils” of the singularities (such fossils being “simpler” languages, “simpler” religions, and “simpler” drawings and cave paintings).

There is a long tradition, inherited by generative grammar from structuralism, of studying languages as formal systems, divorced from their deeper purpose, which is to prompt for elaborate meaning construction (Dancygier & Sweetser Reference Dancygier and Sweetser2005; Hougaard & Oakley Reference Hougaard and Oakley2008; Liddell Reference Liddell2003). In contrast to this tradition, recent research offers very strong evidence for uniform conceptual mapping capacities that cut across domains. The cultural construction of classical mathematics through successive double-scope conceptual blends is demonstrated in detail in Lakoff and Núñez (Reference Lakoff and Núñez2000). Similar detailed research now exists for religion and magic (Slingerland Reference Slingerland2008; Sørensen Reference Sørensen2006); design (Imaz & Benyon Reference Imaz and Benyon2007); technology (Hutchins Reference Hutchins2005; Pereira Reference Pereira2007; Williams Reference Williams2005); poetry, fiction, and theater (Dancygier Reference Dancygier2006); music (Zbikowski Reference Zbikowski2001); social cognition (Turner Reference Turner2001); and, of course, grammar (Fauconnier & Turner Reference Fauconnier, Turner and Goldberg1996; Mandelblit Reference Mandelblit1997). Strikingly, none of these highly typical human domains of action and meaning are possible without double-scope conceptual blending, and their emergence takes place at the beginning of our own cultural time, circa 50 thousand years ago (Klein Reference Klein1999).

We concur with C&C's second principal claim: that language is shaped by the brain. But again, we reach this conclusion by a different route of reasoning. Once the brain, through evolutionary time, acquired the capacity for double-scope blending, language, along with other human singularities, became possible. Language is only one of the universal surface manifestations of the deeper unity of highly human-specific meaning construction, requiring double-scope capacity, and constrained by the constitutive and governing principles of integration (Fauconnier & Turner Reference Fauconnier and Turner2002). Languages can then change through cultural time subject to such constraints, and the well-studied pressures and mechanisms of diachronic linguistics. It remains highly doubtful that such change makes languages better adapted, as C&C wish to think. However even if it did, the capacity for more or less adapted, but fully expressive, languages, attested only in recent cultural time, would remain on their account an unexplained singularity. Accordingly, we view C&C's speculations about linguistic change in cultural time as orthogonal to the central problem of how evolving mental capacities reached a critical point where human singularities, including language, became possible.