of Cognitive Linguistics overlaps significantly with that of philosophy, psychology, and the cognitive sciences, we will concentrate on the history of linguistics only, with an occasional excursion to the history of philosophy. (The chapters of this Handbook devoted to psychology, cognitive science, and philosophy include ref- erences to a number of forerunners in these fields; see Harder, chapter 48, and Sinha, chapter 49.) Section 2 of this chapter briefly describes the internal history of Cognitive Linguistics. The following sections discuss three topic areas of spe- cific importance for Cognitive Linguistics: polysemy, metaphor, and metonymy; the embodiment of cognition; and the Gestalt nature of linguistics. 2. The Short History of Cognitive Linguistics Cognitive Linguistics emerged from its dissatisfaction with dominant orthodox- ies in twentieth-century linguistics, among them the structuralist/formalist tradi- tion in European semantics, the generative/formalist tradition that dominated re- search into syntax in North America, and the formalist/computational approach to semantics that prevailed in North America and Europe during the second half of the twentieth century. Natural allies of Cognitive Linguistics by contrast are func- tionalists and contextualists of all persuasions from the Prague school onward: Functional Grammar (Dik), Systemic-Functional Grammar (Halliday), functional- typological theories of language (Givo ´ n), pragmatics (ordinary language philos- ophy, Grice), Natural Morphology and Natural Phonology (Stampe, Dressler, Donegan), as well as the Columbia School of linguistics with William Diver as its head (who himself followed in the footsteps of Andre ´ Martinet). As Langacker (1998: 1) wrote, ‘‘The movement called Cognitive Linguistics belongs to the func- tionalist tradition.’’ This means that in contrast to formalist approaches, language is no longer viewed as an autonomous system, but rather ‘‘as an integral facet of cognition (not as a separate ‘module’ or ’mental faculty’). Insofar as possible, lin- guistic structure is analyzed in terms of more basic systems and abilities (e.g., perception, attention, categorization) from which it cannot be dissociated.’’ The dissatisfaction with orthodoxies brought with it a questioning of vari- ous assumptions and divisions on which traditional linguistic research was based, in particular the separation of objective knowledge from subjective knowledge, of linguistic knowledge from encyclopedic knowledge, of literal language from fig- urative language, of conceptual/cognitive structures from linguistic structures, and finally of synchronic structures from diachronic change (see Peeters 1998). The in- fluence of prototype theory (and also fuzzy logic) brought about a reevaluation of what had always been put into the formalist-structuralist wastebasket, namely, var- iability, polysemy, and diachronic semantic change. Whereas previous generations of linguists had tended to search for simplicity, monosemy, regularity, and rules, cognitive linguists revel in complexity, flexibility, and patterns, including irregular 590 brigitte nerlich and david d. clarke ones. ‘‘One of the reasons for the emergence of CL and one of its most significant features nowadays is a special interest in those aspects of language that were previously considered as irregular or marginal’’ (Berna ´ rdez 1999: 13). Further, the influence of a new type of cognitive science (that has been called ‘‘Second Generation Cognitive Science’’; see Brockman 2000;Sinha,thisvolume, chapter 49) brought with it a shift from seeing the mind as a disembodied manipu- lation of formal symbols and of language as a syntactic arrangement of formal symbols to seeing mind, meaning, and language as embodied. Syntax, semantics, morphology, and phonology all came to be seen as exploiting universal features of human perception, bodily structure, and social interaction. This means that ‘‘cognition’’ and ‘‘pragmatics’’ are, in a sense, integral components of all aspects of language. The beginnings of Cognitive Linguistics lie somewhere round 1975, which is the year when Lakoff appears to have used the term ‘‘Cognitive Linguistics’’ for the first time (see Peeters 2001). Around that period, Lakoff abandoned his earlier attempts to develop a Generative Semantics by merging Chomsky’s Transformational Gram- mar with formal logic. As Lakoff points out in his interview with Brockman (2000), ‘‘Noam claimed then—and still does, so far as I can tell—that syntax is independent of meaning, context, background knowledge, memory, cognitive processing, com- municative intent, and every aspect of the body.’’ However, in working on his Generative Semantics, Lakoff noticed ‘‘quite a few cases where semantics, context, and other such factors entered into rules governing the syntactic occurrences of phrases and morphemes’’ and caused what generativists saw as ‘‘irregularities.’’ At the same time, Lakoff realized that figures of speech, such as metaphor and me- tonymy, were not just linguistic decorations, or, worse still, deviations, but a part of everyday speech that affects the ways in which we perceive, think, and act. He began his collaboration with the philosopher Mark Johnson in 1979, and they published their seminal book Metaphors We Live By in 1980, which was the first publication to bring Cognitive Linguistics to the attention of a wider audience. But George Lakoff was not the only one dissatisfied with transformational lin- guistics during the 1970s. Typically, ‘‘Cognitive Linguistics has not arisen fully- formed from a single source, it has no central guru and no crystallized formalism’’ (Janda 2000: 3; see also Berna ´ rdez 1999: 11). Around 1975, in fact, Charles Fillmore was working on his theory of frame semantics, and Ronald Langacker was laying the foundations of his Cognitive Grammar (initially called ‘‘Space Grammar’’). Leo- nard Talmy wrote his dissertation in 1972 and began to introduce principles of Ge- stalt psychology into linguistic analysis, especially in his study of force dynamics and event frames (see Talmy 2000a/b). Taking over some of Talmy’s insights into Gestalt psychology, especially the concepts of Figure and Ground, Langacker developed his own theory of conceptual profiling, which became central to Cognitive Linguistics. From 1980 onwards, Cognitive Linguistics began to flourish in the shape of work on metaphorical categorization (Lakoff), image schemata (Johnson), Cognitive Grammar (Langacker), mental spaces and blending (Fauconnier, Turner), and diachronic prototype semantics (Geeraerts). In the second half of the 1980s, Cognitive Linguistics became sociologically organized. In 1989, Rene ´ Dirven, who was particularly instrumental in the international expansion of Cognitive cognitive linguistics and the history of linguistics 591 Linguistics, organized the First International Conference on Cognitive Linguistics in Duisburg, Germany, which became a landmark in Cognitive Linguistics. (Dirven had in fact already organized a ‘‘proto-conference’’ in Trier in 1985.) It was at the Duisburg conference that the International Cognitive Linguistics Association (ICLA) was founded and the journal Cognitive Linguistics, with Dirk Geeraerts as the first editor, and the series Cognitive Linguistic Research, with Rene ´ Dirven and Ronald Langacker (and later also John Taylor) as editors, were launched. During the 1990s, Cognitive Linguistics changed its status from ‘‘revolution- ary’’ to ‘‘established.’’ The biennial conferences of the ICLA that were successively organized in Santa Cruz (1991), Leuven (1993), Albuquerque (1995), Amsterdam (1997), Stockholm (1999), Santa Barbara (2001), Logron ˜ o(2003), and Seoul (2005) witnessed an ever-growing number of attendants, and Cognitive Linguistics may now be said to be one of the major popular frameworks within theoretical lin- guistics at large. There are now also various national cognitive linguistics associ- ations all over the world. In these years of expansion, the historical self-awareness of Cognitive Lin- guistics started to broaden beyond the initial contrastive stance with regard to the immediate competitors, like Generative Linguistics. Some cognitive and historical linguists began to scrutinize the novel (or allegedly novel) concepts used by cog- nitive linguists and discovered that most of them have hidden, forgotten, or scarcely appreciated historical roots (see Geeraerts 1988a, 1988b, 1993a, 1993b; Swiggers 1989; Nerlich 1992, 2000; Nerlich and Clarke 1997, 2000a, 2000b, 2001; Desmet, Geeraerts, and Swiggers 1997;Ja ´ kel 1999). In the course of the following pages, we will illustrate this by looking at three topic areas of specific importance for Cognitive Linguistics: polysemy, metaphor, and metonymy; the embodiment of cognition; and the Ge- stalt nature of linguistics. In each case, we will devote attention to linguists and philosophers who developed theories which can be compared to those developed by cognitive linguists, as well as linguists and philosophers who developed theories which directly foreshadowed and in some instances influenced the development of Cognitive Linguistics. At the same time, we will point to thinkers who developed theories in the more distant philosophical past and those who developed proto- cognitive theories of certain central concepts in the less distant past but were for- gotten in the excitement of the Cognitive Linguistics revolution. 3. Polysemy, Metaphor, and Metonymy It was noticed fairly early on that the closest relative of Cognitive Linguistics in the history of linguistics is probably the tradition of prestructuralist diachronic semantics (see Geeraerts 1988a, 1988b; Nerlich 1992). Figures of speech such 592 brigitte nerlich and david d. clarke as metaphor, metonymy, and synecdoche were not only of interest to philosophers exploring the relation between language and thought, they were in fact also of in- terest to those lexicographers and linguists who were no longer merely looking for the true, original, first, and etymological meaning of words, but came to examine how words were used to make sense by those who used them. They were interested in finding the connections between the meanings of words, in finding patterns in the evolution of meaning, and in putting order into the meanings of lexical en- tries. During the nineteenth century, one can observe a general shift from study- ing meaning as part of etymology to studying meaning as part of a new historical and psychological semantics. Michel Bre ´ al can be regarded as epitomizing this new movement; he was, in fact, also the inventor of a new linguistic term, namely ‘‘polysemy.’’ To get a better idea of the basis for the perceived affinity between Cognitive Linguistics and prestructuralist semantics, we will now have a closer look at Bre ´ al’s work. The ensuing two paragraphs, by contrast, focus on historical links that were misinterpreted or neglected rather than readily recognized by Cognitive Linguistics: Aristotle, and a number of twentieth-century theories of metaphor. 3.1. Bre ´ al and Prestructuralist Semantics From looking at multiple meanings in disembodied lexical entries, Bre ´ al turned to polysemy as a phenomenon of language use, language acquisition, language change, and even neurolinguistics avant la lettre. He wanted to discover the intel- lectual, that is, cognitive, laws of language use and language change (see Bre ´ al 1883). Bre ´ al knew that, diachronically, polysemy stems from the fact that the new meanings or values that words acquire in use (through extension, restriction, metaphor, etc.) do not automatically eliminate the old ones. The new and the old meanings exist in parallel ([1897] 1924: 143–44). And yet, synchronically, or in language use, polysemy does not really exist (it is rather an artifact of lexicographers). In the context of discourse, a word always has one meaning (except in jokes and puns). The most important factor that brings about the multiplication of meanings diachronically and that helps us to ‘‘reduce’’ the multiplicity of meanings synchronically is the context of discourse ([1887] 1991: 156–57). In the constant dialectical give and take between synchrony and diachrony and between meaning and understanding, incremental changes in the meaning of words occur, insofar as hearers, having understood a word in a certain context in a slightly divergent way, become themselves speakers and might use a word in the newly understood way in yet another context, which again brings about a different type of understanding, and so on. In the long run, these slight variations in use and uptake can lead to major semantic changes and, as cognitive linguists have more recently rediscovered, to processes of grammaticalization. More sudden shifts in meaning are brought about by the use of metaphor and metonymy. There are also shifts in meaning which have a more social than poetic root, as when the word operation comes to mean something different according to the social context in which it is used (by a mathematician, a general, a surgeon, and cognitive linguistics and the history of linguistics 593 so on). Analyzing the multiplication of meanings based on the speakers’ and hearers’ social, poetic, and cognitive needs and activities was central to Bre ´ al’s semantics. Bre ´ al was fascinated by the fact that when talking to each other we neither get confused by the multiplicity of meanings that a word can have, some of which are listed in dictionaries of usage, nor are we bothered by the etymological ancestry of a word, traced by historical dictionaries. Both the usage dictionary and the historical dictionary classify the meanings of polysemous words which have been produced over time by a nation or are in use by a nation at a certain time. This is a social (ab- stract and decontextualized) classification, whereas the classification of meanings in the heads of a speaker or hearer is in each case an individual (cognitive, concrete, and contextual) classification. Bre ´ al has in mind an ‘‘isosynchronic competence,’’ a half- conscious type of user knowledge which only works inside concrete situations (see Bre ´ al 1995: 283). It is situated semantic knowledge. Modern polysemy research still debates whether it should predominantly deal with the social or individual type of polysemy and how it should reconcile the one with the other. Bre ´ al observed that most of the time it is the latest, most modern meaning of the word, yesterday’s or today’s meaning, with which we first become familiar ([1884] 1991: 149). Hence, language understanding and language acquisition follow the opposite route of language change; that is, both in language understanding and language acquisition, it is the latest, not the first or primitive meaning of a word, which is the basic meaning. In modern parlance, one would say that the most sa- lient, not the most ‘‘literal’’ meaning, is the one that we acquire first and also use and understand first (see Giora and Gur 2003). Bre ´ al was acutely aware of the fact that the advances made in study of the semantic, cognitive, and developmental aspects of language were not yet on a par with those made in the study of phonetics, of the more physiological side of lan- guage. In his article ‘‘How words are classified in our mind’’ ( [1884] 1991), Bre ´ al therefore appealed to the future to supply us with insights into the cognitive aspects of human language. With Bre ´ al, semantics as a cognitive linguistic discipline made a first step into this future, a future in which we are still participating and to which we are still contributing at the beginning of the twenty-first century—the century of psycholinguistics, Artificial Intelligence, brain scanning, and neuropsychology. Bre ´ al was a central figure in the new tradition of historical semantics inspired by psychology, which had started with Reisig and his interest in metaphor and me- tonymy as mechanisms of semantic change and ended with Stephen Ullmann’s synthesis in the 1960s (see Nerlich 1992). This tradition was resurrected in the light of insights achieved by cognitive linguists in the 1980s with the work of Geeraerts, Traugott, Nerlich, Warren, Koch, Blank, Fritz, Kleparski, and others. 3.2. Aristotle Just like Saussure and Whorf (who will be dealt with further on in this chapter), Aristotle seems to have been the misunderstood whipping-boy of many a cogni- tive linguist even though he ‘‘holds a position of the ubiquity of metaphor 594 brigitte nerlich and david d. clarke in conversation and writing which supports current views about the omnipresence of metaphor in everyday discourse and the print media’’ (Mahon 1999: 69). We do not want to repeat Mahon’s arguments here but only support them with another quote from an early review of Lakoff and Johnson (1980). Here the review er, Mi- chael Smith, argues that Aristotle in the Rhetoric had already remarked that ‘‘strange words simply puzzle us; ordinary words convey only what we know already; it is from metaphor that we can best get hold of something fresh’’ (1982: 128). It is through metaphors that we learn, that we develop our mind and our language. Further nuancing of the picture painted of Aristotle within cognitive linguistic circles can be found in Geeraerts (1989) and Kanellos (1994). It is not this nuanced picture, however, that most cognitive linguists picked up from Aristotle. For them, Aristotle was the originator of two distorted views: an objectivist view of the relation between language and the world and a view of metaphor as simple comparison. Both Lakoff (1987: 157–95) and Johnson (1987: xxi–xxxvi) (see also Lakoff and Johnson 1999: 74–94) argue against the so- called objectivist paradigm in order to then introduce their own so-called non- Aristotelian view of language and cognition. The main tenets of the objectivist or Aristotelian paradigm of thought are that reality is structured independently of human understanding and that this structure is reflected or mirrored in human categorization, where all entities that share a given property or sets of necessary and sufficient properties belong to the same category (this is also called the Classical Theory of Categorization). By contrast, for Lakoff and Johnson, going back to Ludwig Wittgenstein (1953) and Eleanor Rosch (1978), categories are fuzzy, graded, embodied, and changeable, and therefore ‘‘subjective in a nonpejora- tive sense. Yet another part of the Aristotelian straw men that cognitive linguists attacked (see also Richards 1936: 90) was Aristotles’ alleged view that metaphor was purely ornamental. As Mahon (1999, 77–78) has convincingly argued, ‘‘Aris- totle is not claiming that metaphors per se are exceptional. He is only claiming that new good metaphors that are coined by tragedians and epic poets are exceptional.’’ 3.3. Twentieth-Century Metaphor Research At the same time that Lakoff, Johnson, Turner, Ko ¨ vesces, and others were sketching their new theory of metaphor as part of the newly established Cognitive Linguis- tics, there were thinkers in the United States and Europe who, quite independently at first, elaborated their own new theories of metaphor and thought. Their works have parallels with the Cognitive Linguistics research program but are largely ig- nored by it. First, there are a number of scholars who belong to different traditions than the linguistic one—in particular, literary theory, and philosophy. In the United States, Kenneth Burke (1945, 1969) linked the study of rhetoric (the four ‘‘mas- tertropes’’: metaphor, metonymy, synecdoche, and irony) to the study of situated symbolic actions and motives. His work is still very much appreciated by literary cognitive linguistics and the history of linguistics 595 scholars but almost unknown among cognitive linguists (he is mentioned, however, in Turner et al. 1998). In France, the phenomenologist and hermeneu- tician Paul Ricoeur published his seminal book La Me ´ taphore Vive in 1975, in which he discussed conceptions of metaphor from Aristotle up to ordinary language philosophy and tried to bridge the gap between continental hermeneutics and Anglo-American analytical philosophy. An article that Ricoeur wrote for a special issue on metaphor published by Critical Inquiry in 1978, ‘‘The Metaphorical Pro- cess as Cognition, Imagination, and Feeling,’’ was later included in a volume edited by Mark Johnson entitled Philosophical Perspectives on Metaphor, which also con- tained articles by other major European and American philosophers of metaphor (Johnson 1981). Second, some scholars seem to have gone unnoticed because they belong to geographically restricted traditions. In Germany, a whole line of linguists from Jost Trier onwards became interested in studying fields of metaphors, or what is now called ‘‘conceptual metaphors.’’ Trier studied certain domains of experience which constitute major sources for metaphors (bildspendende Felder; cf. Trier 1934: 197–98) and major sources for making sense of the world. Taking up the notion of Bildfeld, Harald Weinrich then developed a theory of metaphor based on the ob- servation of everyday language (see Ja ´ kel 1999: 23). In 1958, he made a distinction between Bildspender and Bildempfa ¨ nger (Weinrich 1976: 284; see also 1967, 1980), which can be compared to that between source and target domain or, as they are sometimes called, donor domain and recipient domain. There are obvious simi- larities between Weinrich’s theory of metaphor and that developed by Bu ¨ hler and Sta ´ hlin at the beginning of the twentieth century (see Nerlich and Clarke 2000a) and the interaction theory of metaphor developed by Max Black in the 1960s—a theory that cognitive linguists did not overlook in their revolutionary fervor (Black 1962; on the relation between the interaction theory of metaphor and modern metaphor studies, see Gibbs 1994). There was one German linguist, who is even less known than Trier or Weinrich and who, in 1954, examined certain domains as sources for metaphors from an on- omasiological perspective: Franz Dornseiff (see Liebert 1995: 149–51). Among many other conceptual metaphors (such as the container metaphor, the metaphor of grasping for understanding, of agitation for anger, and of verticality as an image schema projected onto social hierarchies), Dornseiff discusses what one can call in cognitive linguistic terms the projection of the image schema source-path-goal onto the domain of goal and goal-attainment (see Dornseiff 1954: 142–43; Liebert 1995: 151). Around the same time, the German philosopher Hans Blumenberg published his first essays on metaphor (‘‘Light as a Metaphor for Truth,’’ 1957; ‘‘Paradigms for a Metaphorology,’’ 1960). He had discovered ‘‘metaphor while reconstructing the history of central philosophical and scientific concepts’’ (Ja ´ kel 1999: 23), such as life is a book, which has reemerged as a central metaphor in modern geno- mic discourse (Blumenberg 1986; Nerlich, Dingwall, and Clarke 2002). He thought that the historical study of metaphor could illuminate essential aspects of human 596 brigitte nerlich and david d. clarke existence, culture, and society (see Blumenberg 1997; Adams 1991). As far as we know, neither Weinrich nor Dornseiff nor Blumenberg were ever read by cognitive linguists until they were rediscovered by Ja ´ kel (1999). 4. Gestalt Conceptions of Language Ideas that were originally formulated by Gestalt play a central role in Cognitive Linguistics: foremost among these are the Figure/Ground distinction and, more generally, the idea that meanings do not exist in isolation but have to be understood in a larger context (the idea, in other words, that parts and wholes determine each other). Although Max Wertheimer is credited as the founder of Gestalt theory, the concept of Gestalt was first introduced in contemporary philosophy and psy- chology by Christian von Ehrenfels (see Nerlich and Clarke 1999), and Gestalt psychology developed between 1890 and about 1930. One should be able to find historical affinities between Cognitive Linguistics and earlier linguists who tried to incorporate aspects of Gestalt theory (see Nerlich and Clarke 1999). 4.1. Saussure and Structuralism Although Ronald Langacker informed us (p.c.) that he was not influenced by Saussure’s work in any way and Lakoff never mentions Saussure in his published works, there are some obvious links between Saussurean linguistics and the Cog- nitive Linguistics research program. In order to get a better view of these links, one first has to do away with some common misunderstandings about Saussure (see Nerlich 1999), such as described in the following quotation: He separates individual from society, and language from other non-linguistic sign systems. Society is an anonymous and coercive totality which is external to the individual. The language system is a closed and static system which makes no contact with the world. Saussure is unable to explain variability and change in the linguistic and other signs that we use in social life. Language is a code by which the speaker ‘encodes’ and then transmits non-linguistic ideas and thoughts to the listener in the speech circuit; in turn, these are ‘decoded’ by the lis- tener. The sign is not systematically shaped by its uses in concrete acts of meaning- making. (Thibault 1997: xvii–xviii) These preconceptions or prejudices frequently serve as a backdrop for more modern, dynamic cognitive theories of language and meaning. However, as Thi- bault has shown, Saussure’s views on language can, if interpreted in the context of cognitive linguistics and the history of linguistics 597 sources other than the Cours de linguistique ge ´ ne ´ rale (Saussure 1916), be related to modern theories of schematicity, prototypicality, and indexicality. While this might seem rather far-fetched to some cognitive linguists, it is indisputable that at least some cognitive linguists, like Langacker, share with Saussure a concern with the linguistic sign, ‘‘even when this term is not explicitly used’’ (Thibault 1997: xix; see, e.g., Langacker 1987: 91). At the same time, it must be admitted that there are also clear fault-lines that separate the post-Saussurian structuralist tradition from Cognitive Linguistics, es- pecially insofar as the theorems of the autonomy of the language and the arbitrariness of the sign are concerned. Most interesting for further exploration into the neglected historical parentage of Cognitive Linguistics will, therefore, be those theorists that were inspired by structuralism but that went beyond a static and autonomistic conception of linguistic structure. For instance, in his ‘‘psychomechanics,’’ Gustave Guillaume (1929, 1971) developed a new conception of the language system as a system of systems (similar to Lamb’s stratificational view of language) and the act of speak- ing as constitutive of and dependent on the act of cognition. Karl Bu ¨ hler ( [ 1934] 1990) used Saussure’s distinction between langue and parole to formulate his own prag- matic theory of language, thought, and metaphor. Roman Jakobson (1956a, 1956b) employed Saussure’s distinction between syntagmatic and paradigmatic relationships to formulate a new theory of metaphor and metonymy, of myth and aphasia, and the phenomenologist Maurice Merleau-Ponty ([1945] 1962) developed the dynamic aspects of Saussurean linguistics overlooked by many structuralists. All of them were aware of the developments in Gestalt psychology and all of them anticipated various aspects of modern Cognitive Linguistics. In the following paragraphs, we will discuss the insights of these scholars in more detail; in addition, we will include a section on Whorf, who stressed (in line with Saussure and Humboldt, but, it seems, unaware of their work) that mind without language is essentially amorphous (1956). 4.2. Whorf Lakoff devoted an entire chapter of his famous book Women, Fire, and Dangerous Things (1987) to a refutation of Whorf and the so-called Sapir-Whorf hypothesis, also called the relativity principle, according to which language determines thought (strong version), or, less strongly, according to which ‘‘language affects perception and memory’’ (weaker version) (see Gross 1999: 320). However, several scholars (Ja ´ kel 1999; Stanulewicz 1999; most importantly, Lee 1996) have shown that this ‘‘hypothesis’’ against which cognitive linguists mounted their attack might be noth- ing more than a straw man—similar to the straw man position of Saussure’s view of la langue as an autonomous system. To arrive at a more accurate understand- ing of Whorf, it might be better to abandon the distinction between a weak and a strong version of the Whorf hypothesis (see Brown and Lenneberg 1958) altogether, as it leads to a misleading reduction of a multidimensional problem field in which language, thought, perception, experience, and ‘‘the world’’ interact in 598 brigitte nerlich and david d. clarke various ways. Lee (1996: 27) focuses on this interaction at the interface between language and the world when she writes: When we come to look in detail at the original definitions of ‘the linguistic relativity principle’[,] it will become evident that Whorf’s notion of relativity does not in any way undermine realist acceptance of an independent world be- yond our senses. What it does rely on however, is the understanding that our experiential world (which is the only reality we can say we know) is a function of the human perceptual interface with both the external and internal environment of the human body. Like some interesting recent work in linguistics , Whorf’s experientialism was grounded in insights derived from gestalt theory. According to Whorf, our flux of experience is segmented by culturally, per- ceptually, and bodily grounded patterns of meaning (see Lee 1996: 144). These pat- terns are fluctuating networks of relationships a ` la Lamb: ‘‘Whorf’s mature ideas effectively constitute a field theory of mind in which connections are paramount and entities at any analytic ‘level’ are both indeterminate and functions of the relationships in which they are embedded’’ (Lee 1996: 9). Lee and Stanulewicz point to similarities between Whorf’s thinking and that of various other cognitive lin- guists, such as Lakoff, Johnson, Langacker, and Mark Turner (Turner et al. 1998). Stanulewicz (1999: 193) claims that, just like Whorf, Lakoff ‘‘thinks that the way people use concepts influences the way they understand experience, and be- lieves that differences in conceptual systems significantly influence behaviour.’’ Whorf, just like the cognitive linguists after him, recognized the importance of metaphorical thinking, used the Gestalt concepts of Figure and Ground, dis- cussed image schemas, and saw language as embedded in a network of relations spun between mind, body, and culture. Language has influences on the mind, but language also reflects the conceptual system of the speaker. 4.3. Bu ¨ hler Gestalt psychology also had an influence on Bu ¨ hler’s theory of language, which he developed in the 1930s. It is fundamentally a functional field theory of language (based on the interactions between the symbolic, deictic, and practical fields of language use) which overlaps with a cognitive theory of domains and mental spaces. Unlike Saus- sure, Hjelmslev, or Whorf, Bu ¨ hler developed an explicit theory of metaphor which has some parallels with modern theories of blending. It should be emphasized that Bu ¨ hler’s psychology of metaphor did not appear out of the blue. Its development was prepared by a host of nineteenth- and early-twentieth-century metaphorologists, such as Gustav Gerber, Friedrich Nietzsche, Fritz Mauthner, Gustav Sta ´ hlin (see Nerlich and Clarke 2000a, 2001), and Hans Vaihinger, who wrote that ‘‘all cognition is the perception of one thing through another’’ (Vaihinger 1924: 29). Bu ¨ hler worked in the framework of the Wu ¨ rzburg school of Denkpsychology,or psychology of thought, which had close links with Gestalt psychology (see Nerlich cognitive linguistics and the history of linguistics 599 . theorems of the autonomy of the language and the arbitrariness of the sign are concerned. Most interesting for further exploration into the neglected historical parentage of Cognitive Linguistics. Research At the same time that Lakoff, Johnson, Turner, Ko ¨ vesces, and others were sketching their new theory of metaphor as part of the newly established Cognitive Linguis- tics, there were. interpreted in the context of cognitive linguistics and the history of linguistics 597 sources other than the Cours de linguistique ge ´ ne ´ rale (Saussure 1916), be related to modern theories of schematicity,