In early days, I described blending theory as embryonic. By now, hundreds of people have participated in its development. The commentators in this Forum have suggested further lines of development. It would be an excellent outcome of the Forum if the research community would take up some of these suggestions.

Johnson hits the nail on the head in describing conceptual integration theory as, among other things, an attempt to account for our capacity for abstract thought “without presupposing separate systems allegedly unrelated to our bodily engagement with our environment.” Stjernfelt agrees with the need for an account of abstraction. Danesi reminds us that other thinkers, such as Vico, sought to account for such human mental singularity and, in doing so, offered ideas potentially useful to blending theory. Booth shows us compelling poetic blends and makes the case for poetry as an intensification of imaginative life, in which compressed blends take form. Freeman observes that, in conceptual integration theory, there is every reason to view the arts as having a primary place in the descent of cognitively modern human beings. Yes: art is a great flowering of our species-wide ability for double-scope blending and demonstrates the ways in which advanced blending abilities have endowed us with the capacity to evolve culturally, in cultural time rather than evolutionary time. Human mental operations seem not to have evolved significantly during the last fifty, sixty, seventy thousand years or so, but during that time most of what we regard as marking our humanity has been invented, art often leading the way. Ferynhough indicates the need for better understanding of the ontogenetic development of double-scope ability and remarks that he doubts that I want to strike a nativist note in saying “every human child is born a genius.” “Born a genius?” asks Fernyhough. “Probably so. But I think our best chance of understanding how double-scope blending is possible is to look at how children develop in the first three years or so of life—how those innate endowments are stretched, enhanced, and thoroughly reconfigured by experience out here in the world.” The child is born a genius in entering life with the capacity for advanced blending, but not yet its cultural products. The child deploys the capacity; culture stands ready to feed the child, interactively, the particular conceptual integration networks it has developed. Conceptual integration networks build up—it’s blends all the way down. Often, as in learning numbers, building advanced blends depends upon building others first, with one blend serving as an input to a later blend. The child needs time to progress through such developmental suites.

Pavel remarks, correctly, that it would be unfortunate—as we delve into the way we think—to dismiss everyday understanding of the mind as negligible. “Folk psychology,” sometimes called “commonsense psychology,” is a term of art in cognitive science. It refers to our understanding of human beings as having beliefs, desires, and goals. Perhaps a different term would be better. While there are cognitive scientists and philosophers, such as Paul Churchland, who do argue that folk psychology is profoundly mistaken, folk psychology is already an amazing conceptual achievement—one restricted to human beings, and highly dependent upon double-scope blending in the conceiving of other minds and even one’s own mind. At present, there is no consensus on how to cash folk psychology out into other more sophisticated scientific explanations. Both Herman and Pleshakova, in supplementary ways, emphasize the importance of networks as niches. Because human beings are able to do double-scope integration, they are able to create culture—concepts, artefacts, and behaviors that are not species-wide and not simply induced by variable environmental features but rather conceived in cultural time in certain communities and transmitted to others in the community, chiefly descendents, in ways that can be sustained and developed rapidly even though the earlier members of the community die off. Human beings create and transform robust cultural niches at lightning speed. This is a human singularity: for other species, if there is any culture in this sense at all, it is extremely sparse, fragile, and narrowly tied to basic mental scaffolding. When we cannot get a sure footing with each other mentally, it is not because we do not share the same basic mental operations, but because we do not share all the necessary cultural networks. Luckily, since we are all double-scopers, there is hope even late in life of acquiring cross-niche understanding, by blending. Herman additionally and importantly emphasizes the way in which human thought about the here-and-now involves elaborate conceptual integration networks.

Deal is eloquent in discussing the ways in which blending theory and its analysis of conceptual packing and unpacking offer instruments for analyzing conceptual networks in religion. Harrell makes a crucial observation about future research on such cultural networks and niches: if our identities are imaginative, the result of conceptual blending, then blending theory could presumably help “elucidate the types of ideologies, social relationships, political configurations, and global conflicts that result in our everyday lived experience as humans.” I am particularly grateful for his observation that in imagining the minds of nonhuman animals, we rely on counterfactuality and disanalogy: conceiving of those minds always involves a conceptual integration network that has as one of its inputs our conception of our own mind, with vital relations of counterfactuality and disanalogy helping to structure the network.

Harrell’s proposal to use blending theory to analyze our understanding of identities of other people, other animals, and machines is attractive. Stjernfelt offers a perceptive reminder that the study of double-scope blending must be carried out simultaneously within the frame of human singularity and within the frame of evolutionary descent, and that these two frames are only superficially at odds. The capacity for double-scope blending seems to be species-wide. Its emergence was a fully evolutionary event. That it provided us with the mental capacity for extraordinary discontinuity with other species does not make it any less evolutionary. Stjernfelt also emphasizes that double-scope blending, to be possible, must rely upon a great range of animal capacities that lie in our line of descent. That is a fundamental point: human singularity—in this case, the capacity for double-scope blending—is a small extra step following two billion years of evolutionary descent. But what a difference that small step makes: it gives us the ability to develop robust and inventive conceptual worlds in cultural time, and this is the source of our discontinuity with other species.

Cienki points out a rich area for the study of blending, one that has developed impressively in the last several years—co-speech gesture: “One thing I find fascinating is how not only cultural artefacts (writing, clocks, and other examples discussed in Fauconnier and Turner 2002) reflect conceptual integrations which we employ when we use them, but we ourselves are visibly performing blends any time we use spoken or signed languages.” His survey of the current work on gesture, sign, and blending is an invaluable small introduction. I recommend co-speech gesture as a laboratory to any interested student in cognitive science seeking an area of specialization. It provides many potential dissertation topics—free to a good home.

The commentators raised several topics often discussed in blending theory, and it is good to see them again in the Forum. I comment on a few of them here.

Other animals, “specialness,” hubris, teleology (Rohrer, Herman, and others). When I teach and lecture, I often include riffs about the amazing abilities of other animals. I am frequently cast in the role at conferences and on panels of reminding participants of how weak we are in various ways, mentally, compared to other species and how we simply lack some of their abilities. I explain how, even for mental abilities at which we are superb, we lie on a gradient with other animals, not in a separate galaxy. My sole use of the word “special” in the target article was in my caricature of the cliché objection, “Are we really so special?” “Different” is a better word. The word used in the National Humanities Center initiative underlying On the Human is “singularity,” and that’s the best word, I think, except that it is a technical term. Different people may like different vocabulary, but the facts are not controversial. We have extremely robust culture; cultural evolution is much faster than biological evolution; we are immensely creative in the sense that we invent new concepts and activities that are not species-wide; we conceive over vast scope. These are facts. Blending theory has been explicit from the beginning in reminding audiences that discussion about higher-order human singularities seems to elicit teleological misconceptions about our evolutionary status, and even triumphal misconceptions, dressed up in purple prose, but that these misconceptions must be stopped on the beaches. The capacity for double-scope blending is an evolutionary development. It seems to have been adaptive. Blending theory rejects any triumphal or teleological framing. Indeed, there are many who think that the evolutionary experiment that produced cognitively modern human beings—probably not even a hundred thousand years ago, a blink of the eye in evolutionary time—is doomed to crash and burn, not only for us, but with considerable collateral damage for other species.

The question here is not triumph or teleology but rather our singularities, and it is important to remember that researching human singularities presents special technical challenges, some of which have to do with the place of other animals in comparative research. We are often willing to take drugs because they have been tested on animals. Our reasoning depends upon our belief that, for the relevant systems, the test animals have biology analogous to ours. No doubt, throughout the scientific study of human beings, it is good to look for robust “animal models.” But it is crucial to remember that for higher-order human cognition—art, religion, grammar, mathematical insight, creativity, scientific inquiry, etc.—there are no robust animal models. Beavers, border collies, and barn owls do amazing things, things important to understand, but they do not present us with models for human singularities. All animals exist in networked ecologies in which their actions have long-range consequences, simply in virtue of their having metabolism, reproduction, and local habitation. The beaver’s dam-building is only a particularly visible and memorable instance of such ecological consequentiality. But there isn’t any evidence, for example, that beavers think about and design those extensive and cross-generational networks of consequentiality, or are aware of them mentally as networks, and it’s indisputable that they do not conceive, contemplate, plan, and install inventively different networks in cultural time. We do not have all the abilities that other animals have, and the networks that nonhuman animals inhabit are not fully or even in some cases largely available to our human abilities, but those animals do not have human higher-order cognition. Cats are a great animal model for studying human vision, but they don’t paint, and they don’t think at network scale. We do indeed blow entirely past other animals on the scope of thought.

Neurobiological substrate (Brandt, Benzon, Rohrer). One string of responses to commentaries asks about the neurobiological substrate of blending. I have previously published some stretch speculations about that substrate: (1) Antonio Damasio, in Descartes’ Error, puts forward the notion of “broker neurons,” which might connect up what we think of as disparate neuronal groups subtending clashing conceptual arrays. (2) What is now referred to as the “mirror-neuron craze” has naturally led to notions that mirror neurons (and maybe canonical neurons) subtend blends of self and other. (3) Conceptual integration could be a hypertrophy of perceptual integration: the neural mechanisms of perceptual integration might have been recruited and expanded by biological evolution, resulting in a computational ability that made double-scope blending possible. Perceptual integration, called the “binding problem,” is perhaps the major open scientific question in neuroscience. (4) Synaesthesia, or more generally cross-wiring, could provide neurobiology useful for blending. V. S. Ramachandran, Edward Hubbard, and others have worked on the neurobiology of synaesthesia and considered its contribution to conceptual integrations involving metaphoric links. Stephen Mithen, in The Prehistory of the Mind, has also considered cross-wiring. Synaesthesia is a kind of neural binding in restricted domains. Perhaps it could have evolved into an ability that is not restricted to particular conceptual domains. (5) There are other restricted-domain abilities that look as if they involve integration, such as chase play, a kind of simulation of aggression, which evidently is common throughout the mammalian world for species involved in predation. During chase play, parent and offspring simultaneously activate motor patterns, attention patterns, and motivational structures that belong to two clashing domains, such as parent-offspring and predator-prey. Maybe the neural circuitry subtending binding, synaesthesia, or special-purpose blending of the sort we find in chase play got the ball rolling in the run-up to full cognitive modernity.

Behind the comments about the neurobiological substrate seems to lie an assumption that I do not share, namely that our technology for brain imaging is even remotely close to allowing us to detect double-scope blending as such. I love brain imaging: I am spending this year as part of a three-co-PI team designing and running a set of behavioral experiments, and the entire team has worked hard to force the experiments to conform to the practical limitations on ERP and the extremely severe limitations on fMRI (functional Magnetic Resonance Imaging). We will be running some of those fMRI experiments through the Center for Neuroeconomic Studies. Naturally, we will do what we can, and hope springs eternal. But it’s best in cognitive science not to be carried away with enthusiasm about the latest methods. Non-invasive brain imaging on neurotypicals is only a couple of decades old, and extremely crude. What would we need in the way of brain imaging to be able to detect blending as opposed to other neuronal activity? Consider that blending appears to operate throughout all conceptual domains, and constantly. I take it that very few of the brain’s blending attempts ever advance past the initial stage, very few of those actually conform to the constitutive or optimality principles, very few of those have effect on thought, very few of those have access to action, and only the tiniest fraction are ever accessible to consciousness. The very seductive fMRI images we see in grant proposals are seductive partly because most readers are unaware of how they are produced. They are crude measures of the paramagnetism of relatively deoxygenated hemoglobin. fMRI is a BOLD response. BOLD stands for “Blood-oxygen-level dependent.” Two of the scientists who helped develop MRI were awarded the Nobel Prize, and it’s fantastic for detecting what part of the shoulder the weekend warrior blew out lunging for the tennis ball. But when applied to the brain, it’s still a measure of blood flow, not neuronal activity. It has low signal-to-noise ratios. The results presented in those colored images derive from many repetitions and then statistical averaging and smoothing, often involving wholesale subtraction. There are claims that fMRI correlates better with input than output. At present, fMRI has many exceptionally severe limitations and uncertainties. If we ask, “where does conceptual integration happen in the parts of the neocortex in which fMRI can detect activity?”, the off-the-cuff guess would be “everywhere.” And if we ask, “when does conceptual integration happen in the parts of neocortex in which fMRI can detect activity?”, the off-the-cuff guess would be “all the time.” It is difficult to see, then, how current fMRI techniques could offer any insight into conceptual blending.

Rohrer is on the right track, I expect, in calling for advances in cognitive neurophysiology rather than cognitive neuroanatomy. The neurophysiological processes will be very important; it’s not yet clear to what extent techniques of anatomical localization, even better ones, could help. To be sure, I was delighted when fMRI was developed; it is a useful addition to the arsenal of indirect approaches to mental activity; and we all hope that new and better methods will be invented all the time. In particular, ecologically valid fMRI would be brilliant—measurements taken when people are vibrant in ecologically valid activity, rather than still, silent, alone, supine in a claustrophobic tube, following a white-room experimental protocol, often restrained by soft pads and biting a bar to eliminate motion—but it isn’t easy to be optimistic that new kinds of brain imaging will give us insight into the neuronal substrate of blending any time in the next few decades.

Plus ça change, plus c’est la même chose: When I was studying neurobiology as an undergraduate, back in the early 1970s, the attitude seemed to be that if we learned enough about the plumbing, somehow a theory of thought would precipitate. So we studied ion pumps, myelination, thresholds, and axonal spikes, not to mention neuroanatomy, but never anything like invention, consciousness, the conception of personal identity, the understanding of other minds, or even language. Of course, theory of thought did not easily precipitate from this nuts-and-bolts research, and the vanguard of cognitive neuroscience turned to theorists of mind to learn what neuroscientists should be looking for in all those nuts and bolts. So far, the contributions have been mostly asymmetric, from theory of mental activity (like language) to neuroscience. What would be most welcome would be an avenue along which neuroscience could contribute to the development of blending theory and to specifying the neurobiological evolution that made it possible.

Rohrer writes, “So it might be true that human beings have evolved to be unique in the capacity for double-scope conceptual integration, but if so I want to know how and why primate, dolphin, and other mammalian brains and social structures are not capable of supporting and fostering such integrations, and I want to have a clear story from comparative neurophysiology about how the tree of life yields first something like image schemas, then something like single-scope blends, then full scale double-scope blends.” That would be great, if it exists. But this call reminds me of a passage from Henry IV, Part One:

Glendower: I can call spirits from the vasty deep.
Hotspur: Why, so can I, or so can any man;
But will they come when you do call for them?

Not having the necessary neuroscience, the effective monitoring techniques, or a time machine, it’s not clear yet how we could call such vasty spirits from the evolutionary depths, or even, if we could, what that knowledge would tell us about the operation of conceptual blending.

Explanation and prediction (Cánovas, Rohrer, Brandt, Tobin). The literature in philosophy of science on the nature of explanation and prediction is vast. It is impossible here to say more than a few words. (But see Fauconnier, Gilles, “Methods and Generalizations, in T. Janssen and G. Redeker, editors, Cognitive Linguistics: Foundations, Scope and Methodology. The Hague: Mouton De Gruyter, 1999, pages 95-127.) The central method of scientific explanation is generalization over data: theory—in the form of efficient generalizations—is put forward and is then tested for its broad application to data that were not part of the set used to conceive of the theory. Newton’s laws of motion, for example, fit this characterization of theory, explanation, and prediction. Conceptual integration theory also fits this characterization. For example, in “Rethinking Metaphor,” Fauconnier and I make assertions derived from blending theory, that for the relevant (very large) communities, there is an extensive conceptual integration network for understanding time, and that it contains certain smaller conceptual integration networks organized by certain vital relations, projections, and compressions following the lines of the principles of blending that our theory lays out. We assert that our set of generalizations applies very broadly, efficiently, and usefully to the data. We assert that this theory captures great ranges of conceptions, expressions, and actions by human beings, and that it is a basis for understanding each other, and that human beings will continue to use it in the future very widely for conceptions, expressions, actions, and understandings. We also show that for large categories of these conceptions and expressions, rival theories that do not include blending (e.g., basic metaphor theory for TIME IS SPACE) fail to capture the kinds of data that we have captured. Our theory offers scientific generalizations that are broader and more integrated and that make better predictions over the data. Our theory is better, where better and worse are to be judged according to the standard expectations of application to data. I offered an analogous demonstration in the chapter on “Analogy” in Cognitive Dimensions of Social Science, in which I asserted that for analogy theories that do not include blending, even the examples adduced as the best support for the theories will fail to capture central inferences, and I showed, I believe, that this is true, by taking seriatim any major chestnut examples I could find that analogy theorists had used to support their theories, and showing their inadequacies. This is how science works. Of course, blending is not algorithmic or deterministic. Given that theory must preserve phenomena, a theory that proposes to make algorithmic or deterministic predictions of blending should be shunned.

In the time since Fauconnier and I published The Way We Think, many of the types of conceptual integration networks we put forward—and indeed many of the specific blends—have shown up repeatedly in data that did not even exist at the time. Scientific theory of this sort generalizes but is not reductive in the sense of eliminating one level of phenomena by redescribing it in another—Newton’s laws of motion, for example, do not attempt to reduce motion to some underlying substrate that is not motion. In cognitive science, there are explicit attempts at reduction, such as eliminative materialism. The debate over the wisdom or meaning of such reduction is extensive. Just for starters, although we are all keenly interested in the neurobiological substrate of cognition, it is not clear that finding such substrates constitutes either reduction or explanation. We know, for example, a great deal about the electrochemistry of neurons, but that does not mean that we understand the role of the neuron in thought. It would a splendid advance in scientific knowledge to discover neurophysiology widely underlying blending—indeed, perhaps science has already discovered much of it, but doesn’t understand how that neurophysiology makes blending possible—but it is not clear that this wonderful new knowledge would at present give us insight into the nuanced operations of blending. Maybe, maybe not.

The Human Condition (Pavel, Freeman). Blending theory is not triumphal. It’s easy when reviewing art, poetry, complex numbers, and institutions for decision-making, law, and politics to wax rhapsodic. But human beings seem to suffer greatly from the mental arrays they can construct. They also feel responsibilities, frustrations, and ambitions that are possible only because of the scope of human thought. “Who has twisted us like this?” asks Rilke. “Wer hat uns also umgedreht?”

. . . the shrewd animals
notice that we’re not very much at home
in the world we’ve expounded.
und die findigen Tiere merken es schon,
daß wir nich sehr verläßlich zu Haus sind
in der gedeuteten Welt.

No person, thing, idiosyncratic culture, or local event has twisted us like this, but rather our common phylogenetic development for a mental capacity that brings unprecedented power but no guarantee of pleasure—double-scope blending. Even as it brings the capacity for a sense of responsibility and purpose, guilt and redemption, meaning and value, it also brings a capacity for deliberating over what to do, what to be, how to behave. Pavel emphasizes that we want human life to be not only feasible, but also worth living. “. . . our bio-psychological endowment generates a specifically human ability to live not just according to needs but also to norms and ideals. This ability allows us to decide what kind of leadership we want, discriminate between the various ideals we can pursue, and adhere to the norms that govern our actions. I am convinced that Turner’s notion of double-scope blending deserves to be expanded and adapted to a more vivid sense of human nature. It would help explain how we, human beings, are capable to make the difference between right and wrong, justice and oppression, worthy and unworthy goals.” Freeman equally emphasizes the way in which double-scope blending makes possible systems like the arts, which are “crucial and necessary for all human beings to fully realize the possible scope of human cognition within our own individual consciousnesses.” The human condition is not simple: evolution did not so much make us human as provide us with the mental abilities we need to make ourselves human, an on-going and dynamic process, with hope and uncertainty stretching over the vast scope of human thought.

References

Beattie, G., & Shovelton, H. 1999. Do iconic hand gestures really contribute anything to the semantic information conveyed by speech? An experimental investigation. Semiotica 123: 1-30.

Clark. H. 1973. Space, time, semantics, and the child. In T. E. Moore (ed.), Cognitive development and the acquisition of language, 27–63. New York: Academic Press.

Fauconnier, M. & Turner. M. 2002. The way we think. New York: Basic Books.

Kendon, A. 2004. Gesture: Visible action as utterance. Cambridge: Cambridge UP.

Liddell, S. 2003. Grammar, gesture, and meaning in American Sign Language. Cambridge: Cambridge UP.

McNeill, D. 1992. Hand and mind. Chicago: U. of Chicago Press.

Oakley, T. & Hougaard, A. (eds.). 2008.. Mental spaces in discourse and interaction. Amsterdam: John Benjamins.

Sweetser, E. 2000. Blended spaces and performativity. Cognitive Linguistics 11: 305-333.

Now let’s consider the case of Walt Disney’s Fantasia, a film created to provide visual interpretations of well-known pieces of Western art music. In particular, I’m interested in the first segment, based on Bach’s “Toccata and Fugue in D Minor” and the fourth, based on Stravinsky’s “The Rite of Spring.” As is the case with all the segments, the visual imagery is precisely synchronized to the music. The music itself provides the “human scale” of temporal action in the film.

The imagery for the first segment is of two kinds. For the opening toccata, which is in free time, we see silhouettes and colored shadows of the performing musicians. Let’s just set that aside. It’s the fugue imagery that interests me. This imagery is quasi-abstract. Much of it is just lines and colors and forms in motion. But here and there we see representational imagery, e.g. violin bows, highly stylized mountains, clouds. But the representational imagery is so decontextualized that we cannot attribute any particular scale to it; the only scale evoked by the non-representational imagery is that of the physical scale of the imagery itself. Further, while the imagery appears to exist in 3D space, that space has does not have a well-defined structure. Things move in the space and the virtual camera moves as well, but there’s no sense of direction, of moving toward or away from anything. Motion just happens. It’s a world with form and dimensionality but no scale and no orientation.

But all the motion is synchronized with the music, from beats, to measures, to phrases, to the structure of the whole composition. Because the represented world is quasi-abstract the moving imagery doesn’t provide cues about a temporal structure inherent in that world. There is thus no meaningful distinction between the temporal structure of the world represented in the imagery and the temporal structure of the music itself.

The rather remarkable upshot of all this is that, when watching and listening to the episode, one has the sense that the moving images are being moved by the sound. This sense of causality doesn’t correspond to anything in the represented world; it seems to inhere in the artifact itself, in the movie. If we want to explain this sense of causality, we’ll have to do so in terms of the operations of the nervous system.

The fourth episode is quite different. Rather than using Stravinsky’s own program for the music, Disney provides one of his own, quite different from Stravinsky’s. He chose to represent events occurring in a span of time running from before the origins of life on earth through the extinction of the dinosaurs, all synchronized to roughly a half-hour of music. The visual objects on the screen now represent things in the natural world, and at a wide range of scales. The opening shots move through billions of miles of space; no human eye has ever been at a vantage point from which to see the Milky Way galaxy, which we see in the opening shot. A bit later we’re under water observing the actions of unicellular objects, objects that would be invisible to the naked eye. Most of the action, however, takes place at human scale: cataclysmic weather, dinosaurs moving about and, in the dramatic climax, fighting.

Temporality is complex. Of course there is the music itself, which is, by definition, at human scale. It lasts roughly 30 minutes, while the temporal span that is represented spans a billion or two years. Two techniques are used to “project” that vast span into the 30 minutes of music. My sense is that most of that temporal expanse is simply dropped from view. We get brief scenes of represented action that unfolds on the same scale as the music itself; these brief scenes are presented one after the other, but the instantaneous transitions from one scene to the next skip over unspecified but often quite large intervals of time. The other technique is simply to speed up events without actually dropping any out; this happens in the opening pan and zoom from a POV somewhere out in space to a POV just above the surface of a volcano-covered earth.

Somewhere more or less in the middle of this segment there is a short segment that’s different from the anything else. Every other image in the segment is to be taken more or less literally: had you been in the proper place at the proper time, this is what you could have seen. This short segment is not to be taken literally. It is a metaphor, or whatever.

It runs roughly 3 seconds or so; takes place under water; and shows a creature moving from lower left to upper right. The image we see is of human scale, but . . . At the beginning of the segment the creature is a fish swimming in open water; at the end of the segment the creature has four legs and is walking up an underwater bank. Then the virtual camera assumes a POV that’s just above the surface of the water and we see an amphibian of some sort break the surface. What’s going on?

In a different context one might think of the creature as a shape-shifter. But that interpretation is not available in this context. In the bit of narration that introduces “The Rite of Spring” we’re told: “Science, not art, wrote the scenario of this picture. . . . Finally, after about a billion years, certain fish, more ambitious than the rest, crawled up on land and became the first amphibians.” So, we’re asked to give the entire “Rite of Spring” a naturalistic interpretation. OK. But what naturalistic interpretation do we assign to this three-second sequence in which a fish changes into an amphibian? If you don’t know about biological evolution, you may have a problem producing a suitable intepretation.

If you do know about it, then one interpretation becomes obvious: We’re seeing the evolution of fish into amphibians. In this interpretation, millions of years of evolutionary time has been compressed into roughly three seconds of real time. And a complicated process involving genes, phenotypes, reproduction, and living-in-the-world has become projected onto the plastic deformation of an image moving along a certain trajectory on the screen.

I’ve not attempted to analyze this in terms of blending theory, but I suspect that one critical bit concerns self-identity. On the one had we have the identity that exists among the successive images of the moving creature. What we see on the screen is one object that changes form from one moment to the next, not a succession of different objects that happen lie along the same trajectory on the screen. Just how that identity happens, just how those different forms are fused into one object, that is to be accounted for through the processes of the perceptual system. We need to conceptualize that identity as a particular cognitive object, however, because it corresponds to a rather more abstract identity in the evolutionary process, namely the identity of the process with itself over the evolutionary time scale. The emergent blend of those two identities is very abstract indeed, but that does seem to be what this 3-second segment of film is bringing to awareness.

So, where are we?

In one case we have imagery set to the fugue section of Bach’s “Tocatta and Fugue in D Minor.” Because that imagery is quasi-abstract it lacks any meaningful scale other than the physical scale of the images themselves. The temporal scale is that of the music itself. In the other case we have extensive representational imagery set to Stravinsky’s “The Rite of Spring.” Because that imagery is representational it evokes the spatial and temporal scales appropriate to the world being represented. In some cases those scales are commensurate with those of the sound and the images themselves while in other cases the scales are disparate. That scale information is somehow encoded in the image schemas and the semantic and episodic memory systems where that world knowledge is encoded. As that world knowledge is summoned into mental spaces by the film’s unfolding imagery, the scale information will either match or fail to match perceived physical scale of the imagery and music itself.

And then we have the special case of that 3-second evolution segment, where an interpretive mismatch between a piece of human-scale imagery and the overall interpretive frame forces a conceptual blend the evokes a body of knowledge – evolutionary theory – that is otherwise irrelevant. This is a very sophisticated conceptual trick, and one of a kind well-suited for analysis through blending theory.

And that’s what I find most interesting about the blending model, its use in the close analysis of a particular example.

I’m somewhat more skeptical about the broad claims for “vast conceptual integration networks” that Turner invokes in his discussion of Al Gore’s pale blue dot. It’s not that I doubt the existence of vast networks of conceptual knowledge, not at all. But it’s not clear to me what blending theory says about how they work. In particular, it seems to me that blending theory is rather opaque on the distinction between what happens “on-line” within mental spaces (taken as time-limited constructs) and what exists “off-line” in the more or less permanent structures of the mind. Most of the networked vastness is in those off-line structures fragments of which are “written in” to on-line mental spaces only as needed. Much of the integration is encoded off-line as well.

The nature of those off-line structures is both taken for granted and unexamined in conceptual blending theory. This complaint seems rather similar to Tim Rohrer’s complaint that “conceptual integration theory doesn’t yet have a well worked out theory of how embodied sensorimotor mechanisms are projected into or provide the basis for the mental simulations that underlie conceptual integrations.” It thus seems to me that blending theory is in a situation a bit like that of the cartoon character who walks off the edge of a cliff and keeps on going, walking on nothing but air. What happens when you look down?

As to the continuum, I think it is a danger to make the concept of the existing present now too narrow. It is correct to say, as Mark does, that everything in a certain sense exists in the present now only. This present now is not punctiform, however. We know from phenomenology that the conscious aspect of the present now covers the span of seconds – and the preconscius structure of what lies behind the present now is a larger timeframe, determined by the cyclic metabolism of organisms. Take von Uexküll’s old “functional circle” of animal behavior linking action and perception to constitute the characteristic “Umwelt” of an animal species. This “Umwelt” has a larger extension than the here-and-now – ranging from very small Umwelten in time and space in simple animals and to considerably larger Umwelten in higher animals, especially birds and mammals. So the vast explosion of the range of human thought takes its basis in the Umwelten of higher animals which are, already, quite impressive constructions.

As to the discontinuous, or in any case, very quick jump from animal to human thought, several different proposals are on the market. Mark’s and Gilles’s idea of “double scope blending” is indeed insteresting, as is Tomasello’s “shared intention” hypothesis or Terrence Deacon’s symbolic species proposal. Other classics are tool use, syntax, taboo systems, and much more. The difficulty in comparing such accounts is that they have different bases in semantics, psychology, behavioural studies, grammar, anthropology, etc. The answer, when we once may reach it, will have to take the shape of an explanation which may be translated into the different vocabularies of these sciences so as to decide the validity of each of these compteing proposals, probably rejecting some of them, partially accepting and integrating some others. An important addition to these proposals, I believe, is the human ability of abstraction. Charles Peirce distinguisthed several types of abstraction – one is “prescission”, or attention focusing, highlighting one aspect of a phenomenon and leaving other such aspecs undecided – focusing e.g. on the red color of an object, bracketing its other properties. This abstraction type is obviously mastered by many higher animals which may make association chains of such abstracted properties. Another abstraction type is what Peirce called “hypostatic abstraction” – the jump from “red” to “redness” – taking a property and making it into a new, artificial object of thought which may now be used and investigated in many ways. Linguistic devices for hypostatic abstraction abound – the noun suffixes of “-ity”, “-ness”, “-hood”, “-ship”, sentence constructions like “the fact that ..”, verb conjunctions like the infinitive, the participles, etc. This ability is not in itself linguistic but is vastly enhanced by language. Hypostatic abstraction is the possibility to take any aspect of experience and “step back” and make that aspect into a new object of thought which may then be manipulated by extrapolations and thought experiments. When you have made the hypostatic abstraction of “a lifetime”, you can extrapolate that in time and conceive of far longer timespans, that of groups, societies, the universe. I tend to think this device – hypostatic abstraction plus extrapolation – is central to the special semiotic capabilities of man – even supremely skilled higher animals do not seem to master hypostatic abstraction to any developed degree. As such, it might form part of – or be coextensive with – the double scope blending hypothesis.

In stating that “no nonhuman animal, for example, seems to be able to understand that other animals hold beliefs, or what those beliefs might be. No nonhuman animal seems to be able to wonder what its life might be like if it had done something different ten years ago…,” the article begins with an implicit counterfactual. Just as Fauconnier and Turner assert that understanding the phrase “the beach is safe” arises in relation to the abstract frame for ‘danger’ (we must imagine an alternative beach perhaps with stingrays, riptides, or jellyfish) (Fauconnier & Turner, 2002), in asserting that human cognition is different than animal cognition, we can truly only understand the former in light of a fictitious conceptualization of the latter. Through disanalogy, we are to think: animal epistemologies are not like our own — otherwise we would see animals with all of the human characteristics like the personal identities phenomena noted above, and more. And it is our cognitive apparatuses for integrating concepts, categorizing, etc., that allow us to envision Turner’s counterfactual regretful animal ten years down the line. This ability has political consequences, even involving animals. In highlighting their nonhumanness, we may invoke other abstract frames to understand animals: they can be construed as material resources, perhaps spurring some toward conceptualizing notions of conservation, while others toward mass production and consumption. Animals may be construed as symbolic, taken up in some cultures as characters in trickster tales, while in others as national symbols.

This takes us to the final point of this note. The implications of what the article describes as double-scope bending are cultural, ideological, and technical, the ability to see humans as special and distinct from animals relies on the same as the ability to see one group of humans as different from others. And there are similar political consequences. All of the social ills of stigma, and the phenomenon of celebrating multicultural and gender diversity, or the ability to view other humans as resources and to then exploit those resources, are all imaginative cognitive feats as well — albeit with dire consequences (Harrell, 2009; Santa Ana, 2002). Recognizing cognitive patterns on the basis of linguistic patterns, can help to reveal the imaginative quality of many of the categories that are reified in classification structures, institutional structures, and lived human experience of navigating them (Bowker & Star, 1999). If we see that substantial aspects of our personal identities and divisive human categories are not real in the world, but are instead cognitive feats, the implications are profound. We can construct newer, equitable, empowering, and equally well-grounded alternative conceptions of ourselves others, these new blends can also become socially entrenched. While we may be hardwired for some types of categorization, as the article’s discussion of the entrenchment of writing conceptual systems illustrates, culturally esteemed knowledge structures can be broadly disseminated in society.

A humanistic focus leads to a charge for conceptual blending theory. We can look at the imaginative construction of fundamental problems of human conflict, at the human condition (Turner, 2001). We can see if conceptual blending theory’s constructs such as vital relations, selective projection, double-scope networks, optimality principles, and compression render those problems more human scale and help us to better apprehend them (Fauconnier, 2000; Fauconnier & Turner, 2000). Given Dr. Turner’s argument for the immense power of our double-scope blending ability, one would hope that we could learn to avoid some of our more grave human (double-scoped) conflicts — billions of humans, more alike than different, fighting over subdivisions and privileges, on the pale fiery blue dot that is our home.

Bowker, G. C., & Star, S. L. (1999). Sorting Things Out: Classification and Its Consequences. Cambridge, MA: MIT Press.

Fauconnier, G. (2000). Methods and Generalizations. In T. Janssen & G. Redeker (Eds.), Scope and Foundations of Cognitive Linguistics (pp. 95-127). The Hague: Mouton De Gruyter.

Fauconnier, G., & Turner, M. (2000). Compression and global insight. Cognitive Linguistics.

Fauconnier, G., & Turner, M. (2002). The Way We Think: Conceptual Blending and the Mind’s Hidden Complexities. New York: Basic Books.

Harrell, D. F. (2009). Computational and Cognitive Infrastructures of Stigma: Empowering Identity in Social Computing and Gaming. To appear in Proceedings of the Association for Computing Machinery (ACM) Cognition and Creativity Conference, Berkeley, CA (October).

Lakoff, G. (1987). Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago, IL: University of Chicago Press.

Santa Ana, O. (2002). Brown Tide Rising: Metaphors of Latinos in Contemporary American Public Discourse. Austin: University of Texas Press.

Turner, M. (2001). Cognitive Dimensions of Social Science: The Way We Think About Politics, Economics, Law, and Society. New York, NY: Oxford University Press.

Turner, M. (2003). Double-Scope Stories. In D. Herman (Ed.), Narrative Theory and the Cognitive Sciences. Stanford, CA: CSLI Publications.

What is in some ways more amazing to me is our apparent comfort in residing in these religious worlds, of our willingness to live and die in them—and for them. If double-scope blending is “a plausible model of the way that cognition behaves” (see Vera Tobin’s comment), then religious cognition cannot be exempt from this model. Yet if, for instance, religion and literature are both produced at least in part through a process of double-scope blending, what is the mechanism whereby religious worlds engender real, palpable actions in the world, in a way that literature does not? We might choose to die for the glory of God, but I know of no one who has chosen to die in the name of Great Expectations. How does double-scope blending lead us toward an answer to this question? Or, will we need to look elsewhere? One thought is that double-scope blending is rhetorically powerful in ways that human-scale only ways of thinking are not. Al Gore’s blue dot is rhetorically powerful, and may spur some on to environmental activism, in ways that an image of an endangered fish is not. At least with Gore’s blue dot, there is something majestic about the large scale packed down to human scale. Is there also something emotionally powerful here?

I am admittedly a relative newcomer to the literature on conceptual blending and other aspects of cognitive semantics. But the questions raised by cognitive science cannot be usefully ignored by religious studies. My hope is that we are at the start of a new field of “experimental” or “naturalized” religious studies that will have an impact similar to experimental philosophy on traditional philosophy.

This point deserves repeating.

No doubt that cultural, historical factors as well as discourse and rhetoric purposes should be taken into consideration while constructing conceptual integration networks. This point has already been raised in, e.g. Fauconnier and Turner 2002, 2008; Coulson 2001, 2006; Oakley 1998; Lukes, 2007. “We need to face squarely the far greater complexity of integrations that lie behind observable metaphorical conceptual systems, we need to take into account their cultural history, and we need to account explicitly for the emergent structures they produce, both over cultural time and over individual time … ” (Fauconnier and Turner. 2008. Rethinking Metaphor).
No doubt that accounting for all the factors and purposes mentioned above will lead to richer descriptions and more profound understanding of regularities and generalities. However, I think that being a very productive research method conceptual integration should provide a basis for interdisciplinary research rather than “blend with other theories”. Conceptual Integration describes and explores a basic mental operation inherent only to human beings, a ubiquitous form of cognition, and offers a powerful theoretical framework for investigation of meaning construction and understanding. Conceptual integration offers research tools for the study of backstage cognition, reason, choice, meaning, concept formation and change. Since such tools are vital for exploring world cultures and societies, conceptual integration is viewed as forming an essential part of studies in the social sciences and thus as a theory which can provide a basis for their unification (Turner, 2001) as well as interdisciplinary research.
Cristóbal Pagán Cánovas says: “Explanation, as Vera Tobin points out, is a different thing. The capacity of forgetting differences, of preserving identities and relations across mental spaces for their further manipulation, is elegantly exposed by blending theory. But a full account of what makes this process possible requires coordination and advancement of huge bodies of knowledge from biology to culture”.
I am not in a position to say anything here about bodies of knowledge from biology; however, I believe that conceptual integration can be successfully employed for the analysis of language, cognition and culture in their interrelation. My recent case study of Russian novel metaphors has provided an illustration of the utility of conceptual blending as a tool for cultural analysis through the construction of the multi-scope network ‘werewolves in epaulettes’. This metaphoric structure emerged not long ago to reflect important aspects of the construction of post-Soviet Russian identity. This particular metaphor has been especially interesting for analysis within the conceptual integration framework because it is culturally specific. Initial research suggests that despite the individual presence of the concepts relevant for the construction of this metaphoric network, in other cultures there is no blending of these to produce an emergent structure and the linguistic realization equivalent to the metaphor ‘werewolves in epaulettes’.
The blending analysis reveals the complexity of integrations underlying this metaphorical structure. The analysis of the respective conceptual integration network elucidates the metaphor’s cultural specificity, its rhetorical capacity, its potential in motivating secondary metaphoric structures, the relation of metaphor to past cultural and ideological traditions, and the ideological implications of its usage in the contemporary Russian mass-media discourse. The conceptual blending analysis makes it evident that this metaphoric network becomes a new Russian national construal of the post-Soviet mythologized concept of enemy.
The study of this culturally-specific and highly creative metaphoric structure reveals the significance of accounting for both permanent features of cognition and metaphor’s cultural and discourse history in the construction of metaphoric integration networks. The cultural specificity of input spaces and cultural models’ motivational and ‘adaptation’ capacity have a crucial influence on the cross-space mappings and projections in the network. The case study has demonstrated that in the construction of such culturally specific conceptual integration networks as ‘werewolves in epaulettes’, the overarching goal ‘Achieve Human Scale’ is complemented by a noteworthy sub goal ‘Achieve Cultural Scale’.
The principles of conceptual integration behind the construction of the meaning of ‘werewolves in epaulettes-mythologized enemy’ conspire to achieve the overarching goal ‘Achieve Human Scale’ and its noteworthy sub goals: compress what is diffuse; obtain global insight; strengthen vital relations; come up with a story; go from many to one (Fauconnier and Turner, 2002). My analysis of the Russian culturally specific metaphors suggests the existence of an additional sub goal – ‘achieving a cultural scale’. This sub goal, in my opinion, ensures the activation of inputs and selection of mappings and projections which make the blend culture-friendly. Making the blend more intelligible, blazing and memorable, and useful to us as human beings thinking in culture, this sub goal contributes to the goal ‘Achieve Human Scale’. The importance of this sub goal must not be underestimated, since accounting for ‘achieving cultural scale’ helps to elucidate the metaphor in its full capacity. The analysis has shown that human-friendly and perfectly suited to Russian culture, the concept of ‘werewolf in epaulettes’ is unlikely to be constructed in other cultures, even assuming that they have concepts of werewolf, crime, corruption, etc.
I would like to emphasize that central to the application of conceptual blending as a tool for cultural research is the analysis of cultural models and their capacities. The concept of ‘cultural model’ (discussed in e.g. Sweetser, 1987; Shore, 1996; Coulson, 2001, 2006; see also a related concept of conceptual pattern (frame) in Lukes, 2007) presents a technique based on the analysis of textual evidence or linguistic triggers for various conceptual structures, which accounts explicitly for the cultural, historical and discourse dimensions of the conceptual blending analysis.
Cristóbal Pagán Cánovas points out “Conceptual integration is an important piece in this puzzle, but of course there are many others”. May be it is a key piece in this puzzle after all?

I’ve been having some intellectual difficulties with Turner and Fauconnier’s notion of conceptual integration ever since they advanced a theory of its interaction with evolution and sociobiology in The Way We Think. My reservations center less around the jump between conceptual integration theory as descriptive phenomenology and the science of the neural embodiment per se that I see that Per Aage Brandt and others have raised, particularly when what emerges is another chapter in the incessant monism v. dualism debate. I actually think that as a real-time theory of cognition, conceptual integration theory is a plausible high-level account of mental simulation.

Instead, I am more concerned that this emphasis on the “specialness” of human cognition (and proffering conceptual integration theory as “specialness” par excellance) implies a teleological (and not stochastic) understanding of evolution and its selection mechanisms. By contrast, I think we need to examine more closely the continuity of human cognition with animal cognition–and in my mind this is why it is a serious problem that conceptual integration theory doesn’t yet have a well worked out theory of how embodied sensorimotor mechanisms are projected into or provide the basis for the mental simulations that underlie conceptual integrations. To make such claims carry evolutionary weight, we will need to look at differences in (for example) the neural arbor growth patterns both across time in a single individual, across the species, and across different species using comparative neurophysiological methods–all the while remembering that the environments and practical problems different species are afforded vary.

Right now, I think we have a better evolutionary cross-neurophysiological story about how image schemas evolved and work than a story about how conceptual integration works. For example, Donald Knudsen’s research team has put “glasses” on juvenile v. adult barn owls and then tracked not only the behaviorial components (e.g. successful striking the target) across time, but also examined the resulting neural arbor differences in the visual striate cortex. The behaviorial component suggests that young juvenile owls can learn to hunt accurately both with and without their spectacles, while older juvenile and adult owls cannot. Their neuroanatomical studies showed overlapping “dual” neural arborizations in the visual striate cortex of the young juvenile owls, suggesting to me that owls have something like a neurally embodied source-path-goal image schema that feeds information from the visual system forward for integration with motor action.

When I compared Knudsen’s findings with other neuroanatomical and neuroimaging studies on primates and humans in my chapter on “Image Schemas in the Brain”, it was clear that primates and humans have evolved similar but even more malleable abilities for dynamic neural reorganization–one major difference being that the “dual” or multiple neural arborizations seem to either be latent or grow through the organism’s entire life. I’d like to be able generalize from such evidence as to how conceptual integration might tap into these multiple neural arborizations, but I can’t–yet. I have high hopes that we will be able to do this in the future; but when we do I think we will be showing that yet another supposedly unique human trait isn’t quite as unique as we once believed.

So it might be true that human beings have evolved to be unique in the capacity for double-scope conceptual integration, but if so I want to know how and why primate, dolphin, and other mammalian brains and social structures are not capable of supporting and fostering such integrations, and I want to have a clear story from comparative neurophysiology about how the tree of life yields first something like image schemas, then something like single-scope blends, then full scale double-scope blends.

I began this comment on a cosmic scale; I’d like to end it on an evolutionary time scale. In Galapagos, Kurt Vonnegut’s futuristic narrator is a member of a new species that is descended from homo sapiens, who is struggling to tell the story of how our species died out due to the fact that our “big brains” were ultimately not a successful adaption–they led to poisoning the planet, climate change, etc. So forgive me if the part of me that is a humanist as well as a scientist thinks that there’s a bit of hubris in Mark Turner’s views that double-scope conceptual integration networks are what make humans so “different.” Taken at an this vast time scale, and even if it is true that this capability is part or much of what makes us uniquely human, Vonnegut reminds us not only that we may one day evolve into something else but that we humans are always evolutionarily continuous with the animal kingdom. In other words, we ain’t so special after all.