Metaphor infiltrates ordinary discourse around the world. To take one small example, the singer, and occasional actress, Madonna performed a series of concerts in Chile back in 2008 and was widely criticized for her onstage behavior by local politicians and at least one member of the clergy. Recently retired Catholic Cardinal Jorge Medina commented on these events to his congregation by saying, “The atmosphere in our city is pretty agitated because this woman is visiting and with incredibly shameful behavior provokes a wild and lustful enthusiasm.” “Thoughts of lust, impure thoughts, impure acts, are an offense to God and a dirty stain on our heart.” (https://mog.com/fairportfan/blog/265002).

What did the Cardinal mean by his statements that Madonna’s thoughts and acts were “impure” and that our hearts, or those of his congregation, contained a “dirty stain”? Upon reflection, most people may acknowledge that it is impossible for immaterial entities like thoughts to be physically “impure” or that “a dirty stain” could be imprinted onto a human heart. But people can readily recognize the metaphoric ideas that impurity in thinking and acting refers to immoral behavior, and that a “dirty stain on our heart” describes the potential negative impact that Madonna’s behaviors could have upon people’s souls (at least those in Cardinal Medina’s congregation).

How people interpret words as having metaphoric meaning in context is a challenging question, one that has attracted a great deal of attention within cognitive science over the past 40 years (Gibbs & Colston, 2012). But one can also ask why the Cardinal chose the specific words he used to communicate his metaphoric messages. Why, for instance, do people often speak of immoral and unethical behaviors in terms of dirt or impurities? My claim in this paper is that bodily experience plays a crucial role in motivating why people speak metaphorically in the specific ways they do, and that metaphoric concepts (e.g., IMMORAL THOUGHTS ARE DIRTY objects) do not solely emerge from instances of metaphoric language use.

Consider some common English expressions that reflect an association between dirt and bad behavior or thoughts (Stefanowitsch, 2011: 301):

This regularity in metaphoric discourse about immoral thoughts and acts may be explained in several ways. Linguistic theories often distinguish between the historical origins of word meanings (a “diachronic” matter), and the mental processes that enable contemporary speakers to use and interpret word meanings (a “synchronic” matter). Psychological theories of human cognition draw a similar distinction between explanations of behavior based on historical or evolutionary principles (a matter of “ultimate” causation) and those related to what contemporary individuals know (a matter of “proximate” behavior). According to these similar distinctions, then, English may have evolved in such as way so that people metaphorically associated bad thoughts and behaviors with dirt, perhaps because of their real world experiences with dirt as being bad, but only understand metaphoric uses of terms like “impure thoughts” because they have simply, and blindly, learned that one can talk about bad abstract matters in terms of dirt. On the other hand, though, there may be a close link between the ways words evolved to be used in certain metaphoric ways and people’s contemporary knowledge and understandings of these metaphoric word meanings. Thus, people may understand that there is a tight relationship between their experiences of dirty things being bad and their use of terms like “impure” and “dirty” to metaphorically describe abstract entities such as human thoughts or souls.

I raise this latter possibility of a strong coupling of the historical and the contemporary in explaining the reasons for why the Cardinal spoke of Madonna in the ways he did because of ongoing debates in cognitive science on the motivations for metaphoric thought and language. One of the main developments in the study of metaphor over the last 35 years is that common metaphoric ideas, such as DIRTY IS BAD, are seen as fundamental parts of human cognition, especially in relation to the creation and existence of abstract concepts (Gibbs, 1994, 2008; Kovecses, 2002; Lakoff & Johnson, 1999, 2003). These enduring metaphoric concepts, called “conceptual metaphors,” underlie much of the ways that people think and reason, and partly motivate why speakers refer to abstract ideas and event in very specific metaphoric ways, such as when the Cardinal referred to Madonna as having “impure thoughts,” and engaging in “impure acts.” In fact, the existence of various related linguistic expressions to talk of immoral and unethical thoughts and behavior, as shown above, are viewed as being derived from the contrasting metaphoric concepts CLEAN IS GOOD and DIRTY IS BAD. People do not talk immoral or unethical behaviors in just any random way, but do so for highly motivated reasons stemming from their bodily experiences of the correlations for clean things being good and dirty ones being bad. In this manner, more generally, metaphoric language emerges from bodily experiences, and people’s cognitions about those experiences.

Daniel Sanford’s (2012, also see 2013) article “Metaphors are conceptual schemata that are emergent over tokens us use” offers a different perspective on where conceptual metaphors come from and how people interpret verbal metaphors. His main argument, as the title of his article implies, is that metaphoric schemata are created by people’s frequent use of different metaphoric words and phrases, rather than the other way around. Sanford notes, “the striking feature of metaphor is not the productivity of conceptual metaphor- far from it. Rather the same words and expressions, with the same figurative meanings, are repeated over and over.” According to his usage-based account, every instance of linguistic metaphor is stored in long-term memory, with each example increasing in strength with greater frequency of use, and its proximity to other instances depending on their degree of similarity. More specifically, different figurative expressions that indicate a specific cross-domain mapping give rise to an emergent metaphoric schema. Thus, the expressions “Men are dogs,” “She’s a fox,” “John is a gorilla” all contribute to the creation of the metaphoric conceptual structure PEOPLE ARE ANIMALS. As a person encounters further linguistic metaphors relevant to the PEOPLE ARE ANIMALS mapping, the schema becomes strengthened, leading to speakers more frequent use of certain verbal metaphors, and to listeners easier processing of their figurative meanings. But people can understand expressions like “John is a gorilla” without accessing the underlying conceptual schema, or conceptual metaphor, PEOPLE ARE ANIMALS precisely because its figurative interpretation is encoded as part of the conventional meaning for “gorilla.”

Sanford’s position is one that various metaphor scholars have alluded to in the past as a possible alternative to the popular “conceptual metaphor theory” (CMT) view (Edwards, 1991; Steen, 2008; Wilson, 2011). But Sanford’s analysis is the most complete, partly because it is grounded in several linguistic and psychological principles that are independent of metaphor per se (e.g., exemplar theory, conventionalization, lexical strength). The usage-based account has much appeal given that people’s rich exposure to verbal metaphors, and their cognitive abilities to infer more abstract relations from these leading to the creation of higher-level conceptual schemata, or conceptual metaphors. One of the great ironies about CMT, which originated in cognitive linguistics, is its neglect of the power of language to partly shape people’s orientations to the world, including the strong influence that metaphoric language may have on the ways people ordinarily think (yet often are unaware of this influence- see Thibodeau & Boroditsky, 2011).

Many conventional metaphoric concepts may have surely been influenced by people’s exposure to, and use of, different verbal metaphors, and this fact must clearly be acknowledged and more thoroughly empirically explored in contemporary metaphor research. The rise of naturalistic discourse corpora containing rich systems of metaphor offers one place to further examine what aspects of verbal metaphor people are exposed to and participate in creating higher-order metaphoric concepts (see Deignan, 2006; Stefanowitsch & Gries, 2006). At the same time, some computational models of metaphor have demonstrated the possibility that certain higherorder metaphoric schemata can be inferred given exposure to different conventional metaphoric expressions, at least within certain selected domains such as economics (Mason, 2004).

Still, the idea that people’s use of verbal metaphor is either solely or mostly responsible for the emergence of enduring metaphoric concepts suffers from several major problems. First, why is it that people talk about certain ideas and events in the specific metaphoric ways they do? Why did the Chilean Cardinal employ the phrases “impure thoughts” and “stain on our hearts” when referring to Madonna and her influence on his congregation? As noted above, some metaphor scholars assume that whatever historical motivations enabled certain metaphoric words and expressions to arise in language are not relevant to contemporary speakers’ use of metaphoric language. Thus, today’s speakers have little explicit or tacit understandings of why verbal metaphors often have the meanings they do, simply because these are historically obscure.

Indeed, some metaphoric words and phrases have opaque meanings, yet the important fact is that most do not. The enormous literature from cognitive linguistics describes virtually hundreds of metaphoric concepts, evident in a huge assortment of languages, which are partly motivated, typically in terms of people’s ongoing bodily experiences. The Cardinal’s use of “impure” and “stain” reflect the influence of ordinary experience where people typically view things that are clean as being healthy and good and dirty things to be unhealthy and bad. The theory of “primary metaphor” describes a whole system of correlations in bodily experiences that underlie the use of many verbal metaphors (Grady, 1999). For instance, the primary metaphor MORE IS UP (e.g., “Inflation is up this year”) correlates having more of some objects or substance (i.e., quantity) with seeing the level of those objects or substance rise (i.e., verticality). Primary metaphors include mappings such as INTIMACY IS CLOSENESS (e.g., “We have a close relationship”), DIFFICULTIES ARE BURDENS (e.g., She’s weighed down by responsibilities), DESIRE IS HUNGER (e..g., “I craved fame and fortune”), AFFECTION IS WARTH (e.g., “He was warmly greeted by his hosts”), and KNOWING IS SEEING (e.g., “I see the point of your argument”). In each case, the source domain of the metaphor comes from the body’s sensorimotor system.

In addition to bodily based correlations, metaphoric thought may also emerge because people conceptually understand the similarity or resemblance between two entities or events. When Cardinal Medina said that “the atmosphere in our city is pretty agitated” because of Madonna’s “impure thoughts, impure acts” he is making a comparison between the physical environment of the city with Madonna’s “shameful behavior.” Drawing attention to this resemblance may be achieved by what the Cardinal says, yet he personally must have noticed this prior to his specific speech. In this manner, again, people’s metaphoric talk is motivated by prior metaphoric thinking.

More generally, the claim that people’s conceptual metaphors arise from detecting patterns in metaphoric language ignores the critical starting point that metaphoric thought provides. I have no doubt that people’s frequent use of verbal metaphoric token increases the salience of these primary metaphors, both in terms of their actual bodily correlations, and their continued employment of certain metaphoric language. But a major motivation for where these expressions come from and are continued being use rests with people’s embodied experiences and their detection of similarities or resemblances in the real world. In this manner, metaphoric experience through the body provides the essential grounding for why people speak in the particular metaphoric ways they do.

A wide variety of psycholinguistic and neurolinguistic studies also show that people use and understandings of different embodied metaphors actually recruit bodily based concepts and sensorimotor based areas of the brain (Gibbs & Colston, 2012). Together with the cognitive linguistic analyses of embodied metaphors, the behavioral and neuroscientific evidence strongly implies that embodied activity is not only critical to how certain linguistic tokens come into the language, but why they continue to be employed and understood as being meaningful to contemporary speakers. The usage-based account alone cannot account for any of these data on embodied metaphor use.

A related difficulty with the usage-based account of metaphoric concepts is its inability to explain metaphor in nonlinguistic expression. Much research now documents the extensive manner in which metaphor pervades gesture, art, music, dance, and many other aspects of multimodal performance and material culture (e.g., Cienki & Mueller, 2007; Forceville & Urios-Aparisi, 2009; Gibbs, 2008). Thus, bodily based conceptual metaphors are not just evident in systematic patterns of language use, but within many nonlinguistic domains that both express many similar metaphoric concepts, but ones that also are not directly related to everyday speech. This raises the question- where do these nonlinguistic metaphors come from? Are they independent of linguistic metaphors, or perhaps, as might be argued by the usage-account, derived from verbal metaphoric expressions? Most metaphor scholars maintain that nonlinguistic metaphoric expression are motivated, once more, by recurring patterns of bodily experience, although it is again possible that frequent use of metaphoric tokens in language may have some effect on the creation and continual use of conceptual metaphors in nonlinguistic domains. At the very least, the work on nonlinguistic expression of conceptual metaphors must be explained in some manner by the usage-based account if it is to succeed as an explanatory theory of metaphor.

In a similar manner, one important sources of empirical evidence on embodied conceptual metaphor comes from research in social psychology. Different studies examine the influence of primary metaphoric experience on different social behaviors. For example, there is the widespread set of metaphors suggesting that GOOD IS UP and BAD IS DOWN (e.g., “He is feeling up today,” and “There was a downturn in his luck”). Experiments suggest that these correlations in experience effect different evaluative judgments. Thus, people evaluate positive words faster if these are presented in a higher vertical position on a computer screen and recognize negative words faster if they appear in the lower part of the screen (Meier & Robinson, 2004). People judge a group’s social power to be greater when these judgments are made at the top of a computer screen than when presented in the lower part of the screen (Schubert, 2005).

Finally, participants remember emotionally positive images better when these appeared at the top of a computer screen, with negative images being recalled better when they were seen toward the bottom of the screen (Crawford et al., 2006). These findings are consistent with the idea that people conceive of good and bad as being spatially located along some vertical dimension, a concept that arises from good experiences being upward (e.g., being alive and healthy) and bad ones being downward (e.g., sickness and death).

Furthermore, when people physically engage in certain actions, this also can lead them to adopt metaphoric concepts that influence their social judgments. Having people hold warm, as opposed to cold, cups of coffee, for a few minutes led them to judge another person’s interpersonal traits as being warmer (Williams & Bargh, 2008), a finding that is completely consistent with primary metaphor AFFECTION IS WARMTH. Having people make judgments about people’s behavior in a dirty work area caused them to rate the behavior as more immoral than when the same judgments were made in a clean work area (Schnall, Benton & Harvey., 2008). Similarly, asking people to recall an immoral deed, as opposed to an ethical one, made them more likely to choose an antiseptic wipe as a free gift after the experiment (Zhong & Lilgenquist, 2006). All of these different empirical results are consistent with people’s experiences with the primary metaphors GOOD IS CLEAN and BAD IS DIRTY.

Finally, when asked to determine whether a ficticious person is suitable for a job, people judged job applicants to be better if they were also holding a heavier, rather than lighter, clipboard, (Ackerman, Nocera, & Bargh, 2010), which surely reflects the common idea that IMPORTANCE IS WEIGHT. Having people physically moving backward or forward prompts their recollection of past events or thoughts about future events, respectively (Miles, Nind & Macrae, 2010), results showing the influence of primary metaphors FORWARD IS THE FUTURE and BACK IS THE PAST.

All these different findings from social psychology illustrate the power of bodily-based metaphoric associations in influencing a variety of nonlinguistic behaviors. Once more, it is possible that frequent use of specific linguistic tokens lead to the creation of different metaphoric concepts such as the GOOD IS UP, AFFECTION IS WARMTH, BAD IS DIRTY, IMPORTANCE IS WIEGHT and FUTURE IS FORWARD. But a more direct relation exists in people’s bodily experiences between these different source and target domains, which surely promotes the existence of these metaphoric concepts. People’s experiences with metaphoric language may strengthen the integrity of certain metaphoric concepts, yet it is by no means obvious that these schemata only emerge from language use.

A fourth concern with the usage-based account is its vagueness about the ultimate conceptual schemata that emerges from repeated tokens of use and what role these play in metaphoric thinking and language. For example, consider some expressions in American English used to talk about romantic relationships:

According to CMT, these conventional metaphoric expressions are motivated by A ROMANTIC RELATIONSHIP IS A JOURNEY in which the source domain of JOURNEY arises from our continuous bodily experiences of taking different physical journeys. The mapping of JOURNEY onto ROMANTIC RELATIONSHIPS gives rises to a rich set of inferences or entailments, including the following (Kovecses, 2010: 9).

These entailments underlie the meanings of the various conventional expressions listed above, and together describe some of the rich metaphoric understandings about romantic relationships that people rely on in thinking about and speaking of their relationship experiences. Some empirical research in psycholinguistics shows, in fact, that people can readily judge the acceptable, and unacceptable, inferences that arise from conceptual metaphors and apply these evaluations when interpreting the meanings of metaphoric statements in discourse (Gibbs & Ferreira, 2011).

The question is whether the usage-based account can explain this rich set of metaphoric knowledge through its advocacy of an exemplar view of metaphor. Sanford offers several figures suggesting that different metaphoric tokens may lead to the emergence of higher-order schemata, such as PEOPLE ARE ANIMALS, yet does not specify exactly how this process works. What abstraction process leads one to infer that the use of words “on the rocks,”“dead-end,” “nowhere,” “traveling down,” and “separate ways” necessarily leads to the ROMANTIC RELATIONSHIP ARE JOURNEYS schema? It is also unclear that any summary description of the conceptual schemata, as in ROMANTIC RELATIONSHIPS ARE JOURNEYS, necessarily contains the specific set of entailments that are a fundamental part of people’s rich metaphoric knowledge. How does the emergence process, across metaphoric tokens, lead specifically to these various metaphoric conceptual inferences? More generally, the usage-based account may be unable to infer the specificity of metaphoric inferences without the constraining presence of embodied experience. People know about starting points, paths, obstacles along paths, speed in the manner of travel, direction of travel, success in getting to a destination and so on not from abstracting from linguistic tokens but from their bodily life experiences of taking journeys and using this embodied knowledge to make better sense of concepts such as romantic relationships.

Finally, even if the usage-based account tries to explain the existence of metaphoric schemata, without reference to bodily experiences and other experiential, nonlinguistic factors, it never specifically details what use these schemata are ever put to. Sanford acknowledges that there may be a conceptual structure that governs the creation and use of specific metaphoric phrases. Emergent metaphoric schemas may be accessed to derive certain metaphoric meanings, especially in unusual contexts. But he then states “The repeated use of a word or construction to evoke a particular aspect of a source domain causes the form in question to take on a degree of autonomy from its governing schema.” For example, even if the KNOWING IS SEEING schema is activated when “illuminate” is used to convey metaphoric meaning, “speakers don’t need to use the overall schema to reconstruct anew its metaphoric meaning every time that the word is used or uttered in a figurative sense- the word invokes a direct connection between a particular concept within the source domain (increasing lighting making something more visible) and a particular target (making an idea more easy to understand).”

It appears, then, according to Sanford, that the emergence of metaphoric schemata can have some governing relationship with a linguistic token, but that over time, the linguistic expression will become increasingly dissociated from its higher-order schema. To support this position, Sanford cites evidence from psycholinguistics showing how conventional metaphoric expressions may be understood differently than novel metaphors (Bowdle & Gentner, 2005). Yet none of that research explicitly addressed the question of whether enduring embodied conceptual metaphors are routinely activated as part of people’s interpretation of conventional verbal metaphors. Most of the research on this latter topic suggests that people are, indeed, recruiting embodied metaphors as part of their ordinary use of many conventional expressions, including frequently employed idiomatic phrases (see Gibbs, 2011a; Gibbs & Colston, 2012 for reviews).

The only fallback position that can be argued is that emergent metaphoric schemata are recruited in verbal metaphor use but not necessarily relied upon to infer detailed metaphoric meanings in context. Proving this possibility given the extant psycholinguistic data may be challenging. But more importantly, the usage-based account of emergence implies that higher-order metaphoric concepts may be independent from specific tokens of metaphoric language even if these concepts govern the creation of some verbal metaphors. Once again, any argument along this line fails to consider the dynamical links that are continuously reinforced throughout bodily experience and language use, or, more specifically, between bodily experience, abstract cognition, and linguistic metaphor.

A different perspective on the emergence of conceptual metaphors posits that different variables often seen as completely independent in theories of metaphor may have closer linkages than we imagined and jointly affect metaphoric discourse behaviors (Gibbs, 2011a, b). Different evolutionary/ historical, cultural, social, cognitive, linguistic and neural forces work together as a dynamical system to create in-the-moment, and contextually sensitive metaphoric concepts. These different forces operate along varying time-scales (e.g., slow time scales for evolution down to very fast time scales for brain processes). Most importantly, these various constraints are tightly linked, and affect the dynamics of processing at each level of the system. Thus, evolutionary and historical forces exert some influence, in nonlinear ways, on the operation of faster operating cognitive and neural processes. More generally, metaphoric concepts are not “represented” within any one part of the system, but are emergent properties of the system as a whole, yet may also, as all emergent phenomena, still shape the functional operation of different cultural, linguistic and cognitive sub-systems.

An important implication of this view of metaphoric emergence is that both conceptual metaphors and specific linguistic metaphors are always dynamically linked and cannot be fully characterized without reference to the other systems with which they are connected. Language surely matters to the development and use of higher-order metaphoric concepts, but language alone cannot explain where metaphoric concepts come from, why they continue to be employed in linguistic and nonlinguistic experience, or how language, cognition, and the body are so tightly, and precisely, coupled.

Sanford’s advocacy of EMT highlights the possible influence that language may have on the how people acquire and maintain metaphoric concepts. More attention should be paid within the metaphor community to the ways that language shapes cognition. Nonetheless, language alone cannot be the starting point for how conceptual metaphors come into being, from either a diachronic or synchronic point of view. Most notably, EMT provides no theory of meaning to explain exactly why certain words and phrases have the metaphoric meanings they do, how these historically come into being, or how people, including children, come to acquire an understanding of metaphoric language in the tightly embodied ways they do. In this way, Sanford fails to properly account for the cognitive linguistic and psycholinguistic research demonstrating how much metaphoric language is grounded in recurring patterns of embodied experience that are separate from language and indeed manifest themselves in many nonlinguistic domains as well.

One of the biggest problems with Sanford’s account of EMT, and his criticisms of CMT, is the fact that he limits his discussion to only older, traditional accounts of conceptual metaphor and metaphoric language. Frankly, many critics of CMT appear to have stopped reading the expanding literature on metaphor after the early 1990s (see Gibbs, 2011a for discussion of this concern). But as I have noted in this article, there is a great deal of important linguistic and psychological research that offers critical evidence on the bodily, and neural, grounding of metaphoric thought and language which must be carefully read and acknowledge in any discussion of where and how metaphor came into being (see Gibbs, 2011a, b; Lakoff, 2008, 2012; Lakoff & Nunez, 2000; Johnson, 2008). This more recent work, emerging over 20 years, provides a coherent, neurally plausible theory of meaning and concepts and shows in greater depth why metaphor is not simply a matter of language, but is deeply tied, both historically and contemporarily, to ongoing nonlinguistic experience.