On a Chomskyan postulation in conceptual metaphor theory

1 Introduction

The conceptual metaphor theory championed by George Lakoff and Mark Johnson as an important postulation of the cognitive theory of metaphor is largely built on an “embodied philosophy” intensively elaborated in Lakoff and Johnson’s collaborated book Philosophy in the flesh: The embodied mind and its challenge to Western thought (1999), in which the authors attempt to challenge traditional Western thought by proposing an “embodied mind” or “embodied realism.” As stated by Lakoff and Johnson (1999), the focus of the book is “primarily about the conflict between a priori philosophies and empirical findings in cognitive science” (Lakoff and Johnson 1999: 469). In other words, Lakoff and Johnson see modern linguistics as an arena of the confrontation between traditional philosophies on the strength of a priori assumptions and embodied philosophy based on empirical findings in cognitive science. Evidently, Lakoff and Johnson want to make a difference to the research paradigm at the burgeoning of their theory with the help of novel findings in cognitive science. Therefore, it should be pointed out that, theoretically, Lakoff and Johnson’s embodied philosophy of language is drastically different from Chomsky’s linguistic rationalism. Perhaps what they cannot accept in the latter is mostly Chomsky’s notion that meaning is naturally obtainable through the mastery of grammatical rules in syntax, which removes semantics from the very center of language study. This might explain why Lakoff and Johnson spent a whole chapter of their abovementioned book (See Lakoff and Johnson 1999: 469–512) on anatomizing the philosophy of language underlying Chomsky’s transformational generative grammar, with a concentration on its a priori philosophical assumptions. In that anatomy, they sharply question Chomsky’s approach to dealing with syntax, which is revealed in the statement below:

But from a Chomskyan perspective, “syntax” is constrained in such a way that semantic and pragmatic considerations in principle could not enter into “syntax.” Chomsky’s philosophy requires that his notion of “syntax” take precedence over the distributional generalization criterion. Any distributional generalizations over syntactic elements that require the inclusion of semantics or pragmatics in the statement of the generalization cannot be part of “syntax,” since they are ruled out a priori by Chomsky’s philosophical assumptions. The question here is which is to take precedence, the distributional generalization criterion or Chomsky’s a priori philosophy. (Lakoff and Johnson 1999: 482)

To be frank, this criticism hits Chomsky’s linguistics right on the head, for one of the theory’s major problems lies in that its universal syntax is an a priori metaphysical assumption in nature. This problem makes the key notion of the theory unfalsifiable. In fact, before the chapter-long discussion on Chomsky’s philosophy of language, Lakoff and Johnson (1999) provide a critical retrospect of classical Western philosophies including the thought of Socrates, Plato, Aristotle, Descartes, Kant, and the analytical philosophers. Of all the comments and criticism they make in this respect, the following quoted remark represents an important conclusive statement:

Contemporary linguistics is a philosophically saturated discipline. Many of its founders and best-known practitioners have been trained in ordinary language philosophy, formalist philosophy, formal logic, or some combination of these. Many others, through their university training, have assimilated important philosophical assumptions from these and other traditions.

The assumptions made in those philosophies have found their way into contemporary linguistic theories. The result, we believe, has been a clash between empirical linguistics and a priori philosophical assumptions that, knowingly or not, have been adopted by certain linguistic theorists. It is absolutely vital for linguists to be aware of the effects philosophical theories have had on linguistics. Most important, we need to know whether philosophical assumptions are determining the results reported by linguists as empirical results. In addition, there is the fundamental question of whether it is possible to have a linguistic theory that is sufficiently free of substantial a priori assumptions that its conclusions are not determined in advance. (Lakoff and Johnson 1999: 469)

To Lakoff and Johnson, the philosophical clash between empiricism and rationalism is reflected in linguistics in form of the confrontation between cognitive semantics and transformational generative grammar.

In the same chapter of the book, Lakoff and Johnson (1999) further contend specifically that Chomsky’s philosophy of language is a blended version of Descartes’ philosophy and formalist analytical philosophy. Indeed, Lakoff and Johnson (1999) have made a rather close observation of Chomsky’s theory of language, and their division of the theory into the following two parts is very useful for further investigation:

Chomsky’s theory of language thus comes in two parts. The first part is his a priori philosophical worldview, a blend of Cartesian and formalist philosophy. This is not subject to question or change. It defines a philosophical perspective that he calls “the generative enterprise.” To engage in the enterprise is to accept the worldview. The second part is his specific linguistic theory at a given time, whose details have changed considerably several times over the years. While the linguistic theory is still under construction (as it will be as far into the future as we can see), it is not subject to counterexample, according to the Quine–Duhem thesis, since it is still partial. The generative enterprise, as Chomsky understands it, is a long-term philosophical project defined by an a priori philosophical worldview. (Lakoff and Johnson 1999: 473)

Thus, in the generative enterprise as the first part of his theory, Chomsky’s a priori philosophical worldview enables him to give priority to universal syntax or universal grammar, and most of his subsequent research focuses on testifying to the centrality and universalness of these syntactical structures in all human languages. Interestingly, however, Lakoff and Johnson seem to be doing the same in conceptual metaphor theory, i.e. they similarly give priority to conceptual metaphors and examine how they are realized in natural languages. In other words, the way they deal with metaphor or semantics resembles the way Chomsky dealt with syntax.

In fact, Lakoff and Johnson’s attitude toward Chomsky’s linguistics is rather intriguing in that, on the one hand, Lakoff and Johnson (1999) admit to Chomsky’s influence in helping the “first-generation cognitive scientists” to “become aware of the enormous range of phenomena composing the cognitive unconscious” (Lakoff and Johnson 1999: 471) and his contribution to “helping to bring into cognitive science the idea of the cognitive unconscious as it applies to grammar” (Lakoff and Johnson 1999: 471), while, on the other hand, they also deem it as an important rival theory and thus a target of criticism. Lakoff and Johnson (1999) discuss at great length the conflicts and differences between Chomsky’s linguistics along with its philosophical base and the cognitive theory of metaphor together with the newly proposed embodied philosophy. Thus, instead of continuing to show the great demarcation and inevitable controversies between the abovementioned two theories, the present study purports to demonstrate the commonalities between the two.

One crucial point the two theories share is that they both fall into a certain kind of essentialism. What should be emphasized here is that Lakoff (1987) depicts the notion of essentialism as a type of metaphysical assumption accompanying the objectivist metaphysics that he obviously refutes (Jiang 2016: 523). At the same time, Lakoff (1987) also makes a rather sound definition of essentialism, which may be borrowed for our discussion here:

ESSENTIALISM: Among the properties that things have, some are essential; that is, they are those properties that make the thing what it is, and without which it would not be that kind of thing. Other properties are accidental – that is, they are properties that things happen to have, not properties that capture the essence of the thing. (Lakoff 1987: 161)

Checked through the above definition, both theories contain an essence that forms the essential properties of the thing they respectively describe. For Chomsky, the essence of language is universal grammar or universal syntax, while, for Lakoff and Johnson, the essence of human cognition is conceptual metaphor. Since the essence in either of the two theories is not only paramount but also expresses the most essential property of each, we may call the former syntactical essentialism and the latter metaphorical essentialism. With regard to the essence of language in Chomsky’s transformational generative grammar, Lakoff and Johnson (1999) have made rather lucid statements on it:

Chomsky’s Cartesian philosophy requires that “language” define human nature, that it characterize what separates us from other animals. To do so, the capacity for language must he both universal and innate. If it were not universal, it would not characterize what makes us all human beings. If it were not innate, it would not be part of our essence. We must have an inborn universal “syntax” (called universal grammar) that characterizes humanly possible “languages” in a purely formal, autonomous, disembodied way. This universal “syntax,” shared by all languages, defines the essence of what “language” is.

Chomsky’s use of Cartesian essences is twofold here. The capacity for “language” defines the essence of human nature, and universal “syntax” defines the essence of language. (Lakoff and Johnson 1999: 474–475)

Although Chomsky’s linguistic view represents an amalgam of traditional Western philosophical programs, his essence of language mainly evolves from Cartesian philosophy. The above comments specifically reveal the twofold nature of Chomsky’s syntactical essentialism, namely, language as the essence of human nature and universal syntax as the essence of language. Following this logic, universal syntax could be called the essence of the essence. Thus, we could also say that the essence of Chomsky’s theory is the fundamental deep structure that underlies surface structures. In any case, Lakoff and Johnson have keenly observed the centrality of universalness and innateness in Chomsky’s definition of essence. Lakoff and Johnson (1999) further point out that Chomsky’s twofold essences are philosophically assumed notions, and they thus have the following to say:

In understanding Chomsky’s linguistics, it is crucial to recognize that Chomsky’s philosophical assumptions are paramount. They are taken for granted throughout his work and are not subject to question. Chomskyan linguistics is a philosophical project within a hybrid Cartesian–formalist philosophy. We can see this more clearly by looking at the details of what Chomsky adopted from earlier philosophy. (Lakoff and Johnson 1999: 469–470)

For Lakoff and Johnson, essences in Chomsky’s linguistics are “paramount” philosophical assumptions and “taken for granted throughout his work,” being subject to no question. The diction here in the statement reveals Lakoff and Johnson’s true attitude toward Chomsky’s syntactical essentialism. First and foremost, Lakoff and Johnson define the essences of Chomsky’s linguistic theory as theoretical assumptions. Generally speaking, theoretical “assumptions” are to be tested and should be subject to question, but the situation is different in Chomsky’s theory: he coercively turns his assumptions into a priori rules and principles in the human brain and language. Such logic is taken for granted throughout Chomsky’s linguistic theory. For this reason, Lakoff and Johnson could not accept Chomsky’s syntactical essentialism, for essentialism in Chomsky smacks of a certain kind of determinism. Indeed, Chomskyan linguists should first find biological evidence for language innateness and then prove the ontological existence of universal syntax before attaching such heavy importance to these theoretical assumptions. However, their special philosophical worldview has automatically excluded all possible counterexamples. With such an understanding, Lakoff and Johnson thus make the following conclusive remark on Chomsky’s philosophy of language:

Because of its philosophical status, no empirical finding about natural languages could, in principle, affect this characterization of “syntax” or “language.” Any putative finding suggesting that syntax is not autonomous cannot really be about “syntax” or “language” in Chomsky’s sense, and so must be attributed to some other faculty or theoretical component. Chomsky’s term core grammar applies to what is covered by his theory of “syntax.” Anything outside of Chomskyan “syntax” is outside of “core grammar” and thus not part of what Chomsky’s theory is about. (Lakoff and Johnson 1999: 475)

The criticism here is rather pertinent and creditable, in which the basic logic of Chomsky’s philosophy of language is lucidly depicted. Although Chomsky and his followers were very confident about their logic, it could not be accepted by many other scholars. Their linguistic theory has been criticized by “rank-and-file linguists who are trying to provide complete, thorough, and detailed descriptions of languages from around the world as accurately as they can” (Lakoff and Johnson 1999: 478). These theories do not start with “Chomsky’s philosophical worldview and the accompanying notions of ‘language’ and ‘syntax’” (Lakoff and Johnson 1999: 478). As averred by Lakoff and Johnson, what Chomskyan linguists are seeking is in effect “a purely syntactic essence, a set of parameters shared by all languages and known innately by all normal human beings,” which turns out to be “only a relatively small ‘core’ of formal structures” that exclude many features of most human languages, such as evidential systems, politeness systems, classifier systems, aspectual systems, lexicalization systems, and spatial relations systems (Lakoff and Johnson 1999: 478). With findings in neuroscience, Lakoff and Johnson eventually point out that Chomsky’s theoretical assumptions are biologically impossible:

From the perspective of neuroscience, Chomsky’s idea of “syntax” is physically impossible. As we have seen, a completely autonomous Chomskyan “syntax” cannot take any causally effective input from outside the syntax itself. Such a “syntax” would have to be instantiated in the brain within a neural module, localized or distributed, with no neural input to the module. But this is physically impossible. There is no neural subnetwork in the brain that does not have neural input from other parts of the brain that do very different kinds of things (For a critique from a neuroscientific perspective, see Bi, Edelman 1992, 209ff). (Lakoff and Johnson 1999: 478)

Such comments about the mainstream linguistic paradigm made by Lakoff and Johnson during the last few decades make it more than evident that they aspire to bring fresh blood to contemporary language study.

Interestingly, if we take a close look at the conceptual metaphor theory they hold dear or the embodied realism as its underlying philosophy, essentialism does take an important place in Lakoff and Johnson’s philosophy of language. Since the metaphorical essentialism in the cognitive theory of metaphor has been elaborated and demonstrated in Jiang (2016), here we only reiterate the key points of it. First and foremost, Lakoff and Johnson’s embodied realism seems to “follow the Platonic line of ontological realism, which holds the idea that concepts expressed in a language correspond to real states of things, or affairs that exist independently of language” (Jiang 2016: 522). Through seeing conceptual metaphors or “master tropes” as being fundamental and prerequisite schemas, Lakoff and Johnson presume the ontological existence of a certain number of metaphorical concepts that have priority over other ordinary metaphors. To be more specific, Lakoff and Johnson’s metaphorical essentialism entails three assumptions, as summarized by Jiang (2016:523):

1) A well-structured gestalt structures people’s thought and determines the formation of metaphor.

2) There exist universal categories with ordered hierarchies for one gestalt.

3) There is a coherent system under every conceptual metaphor.

This shows that metaphorical essentialism is deeply rooted in Lakoff and Johnson’s conceptual metaphor theory. Lakoff and Johnson’s criticism of Chomskyan linguistic theory also shows that over-concentration on the essence of a theory might leave out a lot of other important elements, and a similar problem also exists in Lakoff and Johnson’s metaphorical essentialism.

Steeped in metaphorical essentialism, Lakoff and Johnson (2003) seem to believe that these abstract meta-metaphorical constructs — conceptual metaphors — that their theory brought forth are fundamental and thus have the monopoly over the structuring of our thinking and behavior (Jiang 2016; cf. Ding 2010). Thus, Lakoff and Johnson’s definition of conceptual metaphor as the essence of their theoretical construction is very close to Kant’s “category,” which is a priori in nature. This also means that, although Lakoff and Johnson took every effort to build “a linguistic theory that is sufficiently free of substantial a priori assumptions that its conclusions are not determined in advance” (Lakoff and Johnson 1999: 469), their notion of conceptual metaphor theory has eventually developed into a kind of metaphorical essentialism that is metaphysical in nature, for Lakoff (1987) has formally admitted this by contending that the essences in an essentialist theory are metaphysical assumptions (Jiang 2016: 523). As was discussed earlier, such a metaphysical construction is able to find its resemblance in Chomsky’s transformational generative grammar, where the a priori deep structure takes up the center of syntactic study. In other words, Lakoff and Johnson have duplicated a Chomskyan postulation in the cognitive theory of metaphor. Besides the most salient essentialism that the two theories have in common, we are also able to find several other similar features they share, and the following section will be a detailed listing of them.

1) Both theories have a connection to modern cognitive science. As is contended by Lakoff and Johnson (1999), transformational generative grammar and the cognitive theory of metaphor are respectively associated with two generations of cognitive science, with the former the first generation and the latter the second. At the same time, Chomsky also sees the study of the human mind through language as an aspiration of his linguistic inquiry. As averred by Lakoff and Johnson, Chomsky’s linguistics is in line with the findings of first-generation cognitive science, and the cognitive linguistics they championed has become part of second-generation cognitive science (Lakoff and Johnson 1999), although what second-generation cognitive science can provide is still quite limited. To Lakoff and Johnson, the ideas of second-generation cognitive science are dramatically different from Chomskyan linguistic theory:

Second-generation cognitive science is committed to looking at language from the broadest perspective. It includes, for starters, all those things you would have to learn if you were to learn a foreign language: the meanings, the pragmatics, the speech-act constructions, constraints on processing, and on and on. It includes mechanisms of grammatical change – called “grammaticalization” – which typically involve lexical items becoming part of the syntax. This phenomenon inherently crosses the line between the lexicon and syntax. Second-generation cognitive science is committed to studying much more of language than is Chomskyan linguistics. From such a perspective, Chomskyan linguistics studies only a tiny part of language. (Lakoff and Johnson 1999: 479–480)

The biggest difference between the two theoretical patterns lies in that second-generation cognitive science “gives priority to the distributional generalization criterion over any a priori philosophical assumptions” (Lakoff and Johnson 1999: 480). Perhaps one of the biggest contributions of second-generation cognitive science is that it creates a free space for empirical study, and many important findings have thus been discovered ever since its burgeoning in the last century. As admitted by Lakoff and Johnson, cognitive linguistics is the very theory of language that seeks to use these empirical discoveries “to explain as much of language as possible” (Lakoff and Johnson 1999: 491). They also claim that these empirical findings are “at odds with Chomsky’s philosophical worldview on virtually every point” (Lakoff and Johnson 1999: 490).

2) Both theories share a similar logic of universalizing a certain category they bring forth. Analogous to the deep structure in Chomsky’s universal grammar, conceptual metaphors or master tropes for Lakoff and Johnson are also universal or near universal categories (Ding 2010; Jiang 2016; Lakoff and Johnson 1999). In conceptual metaphor theory, metaphor users tend to perceive things in a unidirectional way manipulated by only a few conceptual metaphors. Take the conceptual metaphor more is up as an example. Such metaphorical concepts are defined by Lakoff and Johnson as orientational metaphor. In Lakoff and Johnson’s logic, people naturally believe that “more” means “up” due to their physical experience in life. What they did not take into consideration is that people understand that, on many occasions, our bodily experience or intuition is not necessarily reliable. For this reason, we may encounter in daily speech a number of metaphorical concepts that are counterexamples to more is up, which means that the concept “more” is not necessarily understood in terms of “up.” On many occasions in our life, more is also down, for the more things you have, the heavier you will become. For instance, we observe that, in harvest season, the ears of rice that are rich in rice grains hang downwards in the field, while those that are poor in grains are erect. As conceptual metaphors, more is up and less is down contradict the philosophical phrase “less is more.” In fact, more is up and less is down are meta-metaphorical constructs that are generated for the purpose of research. In other words, they are concepts of a metalanguage for metaphor study rather than metaphors actually used by people in real communicative situations. Compared with “less is more,” more is up and less is down are much more meaningless and lifeless concepts.

3) Major criticisms against Chomsky’s linguistics by former linguists are also applicable to Lakoff and Johnson’s conceptual metaphor theory. Hinzen (2012: 636) summarized three “recent prominent and widespread criticisms” (cf. Christiansen and Chater 2008; Evans and Levinson 2009; Levinson and Evans 2010; Tomasello 2005, 2008) on Chomsky’s universal grammar as follows:

1. UG has no coherent formulation and is indeed unnecessary (Tomasello 2005, 2008).

2. UG is in conflict with biology: it cannot have evolved by standardly accepted Neo-Darwinian evolutionary principles (Christiansen and Chater 2008).

3. There are no linguistic universals: UG is refuted by abundant variation at all levels of linguistic organization, which lies at the heart of human faculty of language (Evans and Levinson 2009; Levinson and Evans 2010).

Interestingly, we can find in the conceptual metaphor theory almost the same deficiencies as in UG, and, to some extent, we can safely replace UG with “conceptual metaphor” in the above statements:

1. CMT (conceptual metaphor theory) has no coherent formulation and is indeed unnecessary.

2. CMT is in conflict with biology.

3. There are no metaphorical universals: Conceptual metaphor is refuted by diverse variation of ordinary metaphorical expressions and unpredictable metaphorical creations, which lies at the heart of metaphor. Thus, there are no metaphorical universals.

Such a replacement at least reflects a certain degree of parallelism in the two theories.

4) The relation between conceptual metaphors and ordinary metaphors resembles that of deep structure and surface structure in Chomsky’s transformational generative grammar. Just as Chomsky distinguishes deep structure from surface structure in syntax, ordinary metaphorical expressions in Lakoff and Johnson’s conceptual metaphor theory are surface level structures, and conceptual metaphors are deep-level structures that underlie them. Regardless of their philosophical ground, the major theoretical structure of the two theories are quite alike. For instance, in the argument is war metaphor, the deep structure is argument is war, and other related metaphorical expressions are surface structures that are generated from it. Let us have a close look at the structure of the conceptual metaphor argument is war:

argument is war

Your claims are indefensible.

He attacked every weak point in my argument.

His criticisms were right on target.

I demolished his argument.

I’ve never won an argument with him.

You disagree? Okay, shoot!

If you use that strategy, he’ll wipe you out.

He shot down all of my arguments. (Lakoff and Johnson 2003: 4)

According to Lakoff and Johnson, people’s understanding of the above exemplified expressions are based on the fundamental metaphorical concept of argument is war. The logic here is, however, rather problematic, as the comment made by Jiang (2016) in this regard says:

There are, however, several problems with these examples in the above statement. To be more specific, the first example could be about basketball, and so is the first half of the second example. The second half of the second example could be about shooting, while the third example could be about construction and so on. Even if all these expressions can be put into the same category as related to war, they are not necessarily reflections of a pre-existing category. Rather, “ARGUMENT IS WAR” is an abstraction of what these expressions have in common, that is, a meta-linguistic construction done by scholars. (Jiang 2016: 525)

Therefore, the true relation between argument is war and the exemplified expressions are fundamentally reversed by Lakoff and Johnson. It is not very hard to notice that, in conceptual metaphor theory, the centrality of conceptual metaphors is endowed by its hypothesized priority and monopoly over other ordinary expressions, i.e. a quasi-relation of the deep-surface structure.

5) Both theories fail to provide sufficient evidence to justify their major theoretical assumptions. As stated earlier, Lakoff and Johnson adopted empirical findings in neuroscience to prove that neuro inputs to the neural module which is deemed as instantiating Chomskyan “syntax” are indispensable, indicating that an autonomous “syntax” is “physically impossible” (Lakoff and Johnson 1999: 480). This proves that the hypothesized essence of Chomskyan linguistics is actually in want of neurological ground. To Chomsky, Lakoff and Johnson’s (1999: 507) evidence for neural “metaboly” and evolution cannot effectively refute the philosophy of innateness, for Chomsky does not imply that “innateness” is instantiated through people’s neural systems. The “universal syntax” championed by Chomskyan linguists is, in reality, a metalinguistic construct summarized by former historical linguists before the formation of transformational generative grammar, the value and importance of which is exaggerated by Chomsky. Similarly, cognitive linguists do not seem to have provided sufficient empirical evidence so far that could justify the biological grounds for conceptual metaphors. Also, empirical findings do not justify the priority and monopoly of these “conceptual metaphors” over other ordinary ones. After all, it is quite untenable to take neural reflection and neural nodes as proof for the existence of conceptual metaphors. One fundamental reason for this theoretical dilemma lies in that conceptual metaphors are in fact meta-metaphorical structures or metalinguistic constructs like the syntactic structures in Chomskyan linguistics, the significance of which is over-exaggerated by Lakoff and Johnson.

6) Similar to their criticism of innateness in Chomsky’s linguistics, the a priori assumption in Lakoff and Johnson’s (1999, 2003 conceptual metaphor theory is so “paramount” that it is taken for granted throughout their works and is not subjected to question (Lakoff and Johnson 1999: 470). Lakoff and Johnson also query Chomsky’s universal grammar and its underlying Cartesian reason and put forward experientialism or embodied philosophy based on Putnam’s internal realism. Epistemologically, however, they fail to explain how an individual’s experience can be apprehended by another mind; rather, they have left the question unsettled in their philosophical framework. Thus, Lakoff and Johnson’s theoretical framework also presupposes a “universal reason” or “near universal reason,” for experientialism is based on individuals’ bodily experience and cannot justify conceptual metaphors, which are transcendental most of the time. In fact, these conceptual metaphors are very much like universal concepts. For example, the abovementioned orientational metaphors like more is up and less is down are explained as being based on the “absolute” physical basis of all human beings. Thus, conceptual metaphors seem to be above individual bodily experiences and are more like a pre-existing Gestalt. In other words, Lakoff and Johnson’s hypothesis of species-specific master tropes is very similar to Chomsky’s hypothesis of universal grammar or syntax. Here, let us have a close look at one of Lakoff and Johnson’s criticisms of the Cartesian method of thinking in Chomsky’s linguistics: “The Method of Introspection. Just by reflecting on our own ideas and the operations of our own minds with care and rigor, we can come to understand the mind accurately and with certainty. No empirical study is necessary” (Lakoff and Johnson 1999: 471). Intriguingly, we can find in Lakoff and Johnson’s metaphor theory almost the same logic. Conceptual metaphors in the cognitive theory of metaphor are as paramount and “unquestionable” as the universal grammar or syntax in Chomsky’s transformational generative grammar.

7) Lakoff and Johnson also criticized Chomsky for frequently making substantial changes to his linguistic theory. Then, what about the situation in conceptual metaphor theory? The most famous, also the most often referred to, conceptual metaphor, argument is war, is revised into “argument is struggle” in the second edition of Metaphors we live by, published in the year 2003. Since the difference between argument is war and argument is struggle is so evident, we have great doubts about the fundamentality and universality of their original concept “argument is war.” Why do they have to change it if it is so fundamental and important for other metaphors in their proposed “coherent” system of metaphorical concepts? Most importantly, who has the right to change it if it is really one of the “metaphors we live by”? Besides, they also seem to have abandoned the term “experientialism” in their later writings and substituted it with “embodied philosophy,” because the word “experientialism” never appeared in either Philosophy in the flesh in 1999 or the afterword of the second edition of Metaphors we live by in 2003 (other parts of the book remain unchanged in the second edition). The problem is that they do not mention a word on the internal relationship between “embodied philosophy” and the earlier concept “experientialism” or why they were determined to substitute the latter with the former.

8) Lakoff and Johnson have inverted the relationship between metaphorical expressions and their meta-metaphorical construct (conceptual metaphors), just as Chomsky inverted the relation between surface structure and deep structure. And Lakoff and Johnson’s version of metaphorical concepts might only exist in the minds of linguists. They believe that metaphorical language is a reflection of metaphorical concepts just as Chomsky believes that language is a realization of universal grammar. Let us have a thorough look at Lakoff and Johnson’s inverted logic:

The Love Is A Journey example reveals the fallacy in tenet I clearly. If metaphor were just a matter of words, then each different linguistic expression should be a different metaphor. Thus, each of the example sentences should be entirely different metaphors, with nothing in common among them. “Our relationship has hit a dead-end street” should be distinct from and unrelated to “Our relationship is spinning its wheels,” which in turn should be different from and unrelated to “We’re going in different directions” and “Our relationship is at a crossroads,” and so on. But these are not simply distinct, different, and unrelated metaphorical expressions. They are all instances of a single conceptual metaphor, namely, Love Is A Journey, which is characterized by the conceptual cross-domain mapping stated in Chapter 5. There is one conceptual metaphor here, not dozens of unrelated linguistic expressions that happen to be used metaphorically. Metaphor is centrally a matter of thought, not just words. Metaphorical language is a reflection of metaphorical thought. Metaphorical thought, in the form of cross-domain mappings is primary; metaphorical language is secondary. (Lakoff and Johnson 1999: 123)

In fact, the above statement is rather untenable and, to some extent, misleading. First of all, traditional metaphor theory already recognized the similarity between two situations. Second, nobody denies that there are commonalities among certain metaphorical expressions. Thirdly, metaphorical expressions like “Our relationship has hit a dead-end street,” “Our relationship is spinning its wheels,” “We’re going in different directions,” and “Our relationship is at a crossroads” are not instances of a pre-existing conceptual metaphor; rather, the latter is the result of a generalization process. It is thus very likely that metaphorical thought is not realized through the so-called cross-domain mapping, for Glucksberg and McGlone (1999) proved that there is no such love–journey mapping in our mind when we are using related metaphorical expressions. This means that cognitive linguists have a lot more to do to justify the psychological basis of conceptual metaphors.

This listing might continue, yet the above arguments amply demonstrate that Lakoff and Johnson’s cognitive theory of metaphor resembles Chomsky’s logic and that their notion of conceptual metaphors is very much a Chomskyan postulation. Although both schools might not embrace this kind of comparison, they do have much in common with regard to the way they developed their theories. And the following words stemming from a further reflection of the above two theories could serve as my concluding remark. In fact, when people are speaking or talking in their mother tongue, they seldom refer to grammatical rules in the moment of uttering.

Traditional grammarians’ work is to make observations on the existing language data, and few of them have ever attempted, not even with the slightest effort, to claim the existence of a universal grammar in which daily utterances are rooted. Indeed, to propose an a priori universal grammar means, to some extent, shutting the door to language innovation. After all, the notion of universal grammar never seems to appear in the mind of any traditional grammarian. In fact, the major distinction between traditional grammarians and transformational generative linguists lies in that the former adopted the method of pure induction, while the latter employed the former’s inductive results as their starting point of deduction. In other words, the latter (represented by Chomsky) initiated their deduction by dint of the former’s inductive results.

People in a certain speech community can be rather alert or sensitive to non-grammatical expressions produced by children or foreigners. This means that grammar has certain patterns based on which people can recognize and rectify unacceptable ones in their language. This is, however, a different story in the case of metaphor. People hear or read much fewer “incorrect metaphors” than inaccurate plain sentences. In other words, when using language in an everyday setting, speakers often make grammatical mistakes, but it is harder for speakers to produce actual “wrong” metaphors. Even if people hear someone produce an inappropriate metaphor, they can hardly replace it with an accurate one at once, for there are only appropriate or inappropriate metaphors rather than standard or nonstandard metaphors. Perhaps the reason here lies in that metaphor is language performance on a higher level. In other words, if Chomsky’s universal grammar has more or less acquired its superficial explanatory power by dint of some grammatical coincidences in different languages, Lakoff and Johnson’s conceptual metaphor theory is more likely to lose its universality and rationality due to the abductive and heterogeneous nature of metaphorization.