Fonosemantinės sakmės Lietuvių kalbos žodyne

ALGIRDAS SAUDARGAS European Parliament Committee on Industry, Research and Energy Research areas: cognitive linguistics, cognitive neuropsychology, biophysics. PHONOSEMANTIC SAGAS IN THE DICTIONARY OF THE LITHUANIAN LANGUAGE Phonosemantic Sagas in the Dictionary of the Lithuanian Language ABSTRACT The article presents an approach to explaining phonosemantic phenomena in Lithuanian. This approach divides the content of word meaning into a conceptual part and those contents that express relations in the dynamics of space, time, and force. From this perspective, phonosemantic phenomena are explained as extensions of the cognitive topology of a word-meaning network to phonologically similar but unrelated words. Several examples are examined. The meaning networks of Lithuanian words with K/L; N/R; M/L <> L/M in the root are analyzed. KEYWORDS: phonosemantic saga, reference frame, image schema, force dynamics, cognitive topology. ANNOTATION An account explaining fonosemantic phenomena in the Lithuanian language is presented. This account partitions the content of the word sense into the conceptual part and the content which expresses space, time and force dynamic (STD) aspects. Fonosemantic phenomena under this account are explained as the extension of the cognitive topology of the network of word senses to phonologically similar though not genetically related words. Several examples are investigated. The network of senses of Lithuanian words with consonants K/L; N/R; M/L <> L/M is analysed. KEYWORDS: fonosemantic saga, frame of reference, image schema, force dynamics, cognitive topology. Įžvalgos / Insights 229 ALGIRDAS SAUDARGAS 0. INTRODUCTION “...one’s own thinking can be expressed only in one’s own language.” Antanas Maceina More than three decades ago, Algirdas Patackas’s reflections on Baltic culture appeared in samizdat Saulė ir kryžius (Patackas 1983; Patackas 2014: 30–85). There, the “law of symmetry” was mentioned for the first time: it was observed that when the root consonants of certain words are switched, their meaning also acquires a certain kind of symmetrical meaning, producing a peculiar inversion of meaning (for example: meilė – laimė) (Patackas 2014: 78). Another observation was that the meanings of words with the same root consonants but different vowels remain unified by a common idea, even though the words themselves are entirely unrelated (for example: nokti – nykti) (Patackas et al. 1984; Patackas et al. 2014: 93). From the perspective of formal linguistics, these examples seemed to be unjustified extensions of well-known linguistic phenomena—metathesis and apophony—and could not be accepted. I was intrigued by the linguistic problems raised in Patackas’s study and encouraged to set down my thoughts in writing. That essay (Saudargas 1986) is full of philosophical musings that today might interest only researchers of the samizdat genre. Yet it is founded on an idea that is also at the heart of this article. My attention turned to words whose roots are phonologically similar but whose relationship is unconfirmed. I attempted to reveal their semantic connections. My guide was (and still is) Antanas Maceina’s view of the mother tongue: “The richness of the Lithuanian dictionary and the diversity of its grammatical structure show that Lithuanian thought is capable of penetrating the object of thought to a remarkably broad and profound extent, and of expressing even the barely perceptible yet real nuances of its connections and relations. In this respect, the Lithuanian language is inferior to no other language in which, over the course of history, a high-level philosophy has been or is being created. Our language is an open path to our philosophy. Whether we take it or turn onto the ossified path of foreignisms depends on our insight that one’s own thinking can be expressed only in one’s own language” (Maceina 1994: 309). In that essay, I discussed verbs denoting the sense of sight: žiūrėti, domėtis, regėti, matyti and vysti (vydo), and compared their semantics with those of words having phonologically similar roots. When I was studying these phonosemantic relations, the emerging studies in cognitive linguistics were unknown to me, since under the conditions of the time they were difficult to access, or perhaps entirely inaccessible. I returned to these reflections a decade ago and soon found that cognitive linguistics had the research techniques I needed. At the time, I was also unaware of many other things that I mention today. 230 Acta Linguistica Lithuanica LXXIV Fonosemantinės šaknės lietuvių kalbos žodyne The texts collected in this article were written over the past decade. They were reconsidered and rewritten several times. When bringing them together in one volume, I reviewed, corrected, and expanded everything once again. Much had to be written entirely anew. These texts are of two kinds. The first kind consists of empirical studies in which I sought to be as objective as possible. I therefore defined the objects of study precisely: these are the lexemes in Lietuvių kalbos žodyno (hereafter LKŽ), and those alone. The study aims to describe the semantic relations of phonologically similar words (those whose roots contain the same pair of consonants). Its purpose is to show that some phonologically similar Lithuanian roots (so-called root clusters) also form semantic wholes, which, for convenience, are called phonosemantic tales. The essence of the study lies in breaking down the meaning content of Lithuanian words into conceptual and spatiotemporal components. Spatiotemporal content also includes the interaction of forces (dynamics), so I use the abbreviation ELD (space, time, dynamics). For semantic analysis, I use the techniques of cognitive semantics. I have tried to rely on the most reliable and tested methods—for example, to describe the interaction of forces, I use Leonardo Talmy’s force-dynamic schemas, which have already become classics. All the empirical results are sufficiently evident, although, as I have experienced myself, recognizing some of them requires a certain familiarity with the techniques used. The second kind of text is subjective. It was needed because the empirical studies mentioned above have no place in the traditional spectrum of linguistic studies. Of course, in their free time everyone can explore the meanings of words as much as they wish. Whether such explorations are reflected in poetry is for the reader to decide. If one wishes to bring them to the attention of the scholarly community, they must find a place of their own. In searching for such a place, it became clear that these studies in linguistics are extrasystemic or even extralinguistic. More precisely, phonosemantic tales are an object at the intersection of disciplines, although linguistics, understood in the broadest possible sense, ought to take an interest in them. While looking for analogies, I wandered between mathematical topology and medieval alchemy. In search of arguments, I leafed through monographs, dissertations, and articles from a wide range of fields: from phonosemantics, considered marginal, to contemporary brain research. Such wanderings constantly threaten to turn into aimless distraction of thought, since it is possible to gain an in-depth understanding of fields so far removed from one another only to a very limited extent. On the other hand, paradigms change and new currents take shape only at the intersection of disciplines. For that reason, these texts draw on a great variety of authors and sources, all of which would take several lifetimes to study thoroughly; the references provided, however, will save the curious reader time. It goes without saying that texts of this kind reflect my subjective choices and personal judgments and cannot claim to offer a scholarly, comprehensive survey of any field. Now it is time to move on to more specific insights. But first of all, we must remember the predecessors with whom it all began. Įžvalgos / Insights 231 ALGIRDAS SAUDARGAS 1. INSIGHTS INTO THE METHOD 1.1. Forerunners and Their Critics “Even honey seems bitter to him” (LKŽ) Together with Aleksandras Žarskus, Algirdas Patackas developed the initial insights into a distinctive study of the management of spiritual transformations in ancient Baltic culture, grounded in abundant ethnographic sources (Patackas, Žarskus 2013). Dainius Razauskas assessed this research as follows: “The work of Algirdas Patackas and Aleksandras Žarskus is one of the most interesting, one of the deepest phenomena in contemporary Lithuanian verbal culture, perhaps comparable in spirit to the phenomenon of Vydūnas a hundred years ago” (Razauskas 2006: 47). However, according to Razauskas, “the ‘honey’ of the spirit, gathered through this immense labor, is poisoned by a ‘spoonful of tar’” (ibid.: 45). This is Patackas and Žarskus’s “relationship with language, or rather with linguistics, with the sciences of language” (ibid.). Indeed, a decade earlier, the reflections of Patackas and Žarskus had drawn the sharp (and, apparently, justified) criticism of linguist Simas Karaliūnas (1995). The official world of linguistics saw in the aforementioned studies no phenomenon that could be investigated. Softening Karaliūnas’s criticism, in an impassioned paper read in 1996 and titled “Linguistic Madness and Sacred Language,” Razauskas asserts: “The aim here is not to prove the origin of one word or another, nor the etymological kinship of words, but to draw attention to sometimes truly astonishing correspondences—for example, between the phonetic and semantic symmetry of word roots—while leaving aside altogether the question of the origin of words” (Razauskas 1997: 10). Nevertheless, for some reason Razauskas resigns himself: “In any case, such an aim does not, it seems, really belong to linguistics. In short: the relationship with language under consideration is not linguistics, not science” (ibid.). Razauskas then redirects the discussion toward the phenomenon of anagrams discovered by Ferdinand de Saussure and rediscovered by Vladimir Toporov. He formulates the problem correctly: “Is it possible—and if so, on what basis, with what justification—to link the simplest phonetic symmetry of the consonants in a word root to the semantics of the corresponding words, let alone to an entire, seemingly imagined worldview?” (ibid.). Unfortunately, immersed in the fascinating but still rather obscure current of research into anagrams and the Indo-European poetics associated with them, he emerges with an answer that is distinctly ethnosophical (to use his own terms1): “Thus, we have not only two languages—the ordinary, everyday one, and the sacred, poetic one—but correspondingly two linguistics,” 1 Cf. Razauskas 2006. 232 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne FIG. 1. Academician Paulius Slavenas’s phonosemantic network Įžvalgos / Insights 233 ALGIRDAS SAUDARGAS that is, two different relationships with language, grounded in different theoretical and methodological ways, as reflections of different relationships with reality in general—namely, ordinary, profane linguistics and secret, sacred linguistics” (ibid.: 16). Another forerunner, the patriarch of Lithuanian astronomy, Academician Paulius Slavenas, published a series of articles between 1974 and 1985 in which he studied the connections between sound and meaning in Lithuanian with great interest (Slavenas 2001; 1984). He observed that vowel changes in a word root very often do not alter its meaning substantially, but only modify it. After reviewing various combinations and rejecting unsuitable ones, he drew a diagram on a plane. By his own account, “when drawn not very finely, the diagram occupied about a third of a square meter” (Slavenas 1984: 34). Since Slavenas’s study attracted no attention from linguists, Figure 1 reproduces his phonosemantic network diagram (Slavenas 2001: 278). The diagram itself is merely a way of organizing linguistic information. The consonant structure is connected through the simplest combinatorial procedure. Through his research, Slavenas seems to suggest: let us temporarily forget everything we know about word roots and consider what associations of meaning are evoked by phonologically similar words. Certain regularities can then be discerned in a carefully constructed phonosemantic network. This methodology is common not only in experimental natural sciences but also in theoretical inquiry. To be fair, it should be noted that linguists themselves often use such a methodology. Thus, this kind of experimentation with word roots violates no known laws of linguistics. In their leisure time, anyone is free to enjoy such experiments. But if we wish to contribute something to scholarship, we must draw conclusions from these experiments and defend them. Academician Slavenas draws truly bold conclusions from his research, which do not seem sufficiently well-founded to the linguistic community. I will quote a few passages from the article “The Meaningfulness of the Proto-Language.” “The ancient proto-language of an isolated tribe developed solely through the older generation’s example and genetically programmed speech habits. It consisted not of words (in the sense of the Lithuanian language), but of shorter, polysemous formants, which could later become either word roots or root affixes. More precisely, a formant does not have meanings, but a broader conception, a somewhat vague category of thought; meanings arise only in combinations of formants. A formant is neither a word in the usual grammatical sense nor a unit of a part of speech: in various combinations it could have the meaning of a noun, verb, or other category—that is, it could undergo what present-day English calls conversion; it could serve as a name, nickname, statement, action, or for some other purpose. A single formant gave only a hint of what was being discussed, while its meaning became clear in combination” (Slavenas 2001: 261). “The various observations made here give rise to the rather paradoxical idea that the divergence of the Indo-European languages means a departure not from a hypothetical Proto-Sanskritic proto-language, but from the structure of an archaic, still living Lithuanian language” (ibid.: 273). 234 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne These insights are indeed profound and interesting, but, although they greatly flatter the self-esteem of the Lithuanian-speaking community, they are not sufficiently substantiated. Therefore, linguists’ neglect of the old academic’s research is entirely understandable. 1.2. The Pharmacy of the Mind “The pharmacy of the mind is beneath that roof. Perhaps to cheer a person with an opinion.” Dionizas Poška “The unconscious mind proposes; the conscious mind selects.” Stanislas Dehaene Let us return to Razauskas’s conclusion that “we have [...] two different relationships with language, grounded in different theories and methodologies, as reflections of different relationships with reality in general—namely, ordinary, profane linguistics and secret, sacred linguistics.” Where should we look for a methodology to study the phenomenon that interests us? The first step could be the claim that “everyday, ordinary, profane” linguistics studies consciously perceived linguistic phenomena, while “secret, sacred, poetic” linguistics seeks to reveal what is unconscious in language, that is, directly inaccessible to consciousness. That everything depends on one’s point of view is clear from a simple example. Grammar is undoubtedly a traditional field of linguistics. Yet we can look at the same grammar from two perspectives. Preschool children speak grammatically correctly, although they have never studied grammar. When they begin learning grammar at school, many find this subject tedious and boring. This is well known and clear. Just as when riding a bicycle or driving a car we perform some actions automatically and others consciously, so when speaking we automatically utter words and grammatically correct sentences, while consciousness observes and guides the meaningful context. The word is the window through which consciousness looks onto meaning, but meaning is immersed in the human subconscious, and the boundary between the conscious and the unconscious is not easily graspedconsciously unconscious... Įžvalgos / Insights 235 ALGIRDAS SAUDARGAS concept. In the next section I ask what neurophysiological measurements of brain activity can offer us today. The final sections of these reflections on method define the phonosemantic saga—the subject of this article’s research. A particularly important dimension of language lies between unambiguity and ambiguity (Eng. ambiguity2. It might seem that unambiguity should be the ideal of language. That is indeed the case when ambiguity can be dangerous. An officer’s order must be understood unambiguously. A surgeon’s instructions during an operation cannot be ambiguous. The lives of soldiers or patients depend on it. In legislation and many other fields, the aim is for the terms used to be understood unambiguously. So why is there ambiguity in language? Ambiguity in language is an empirical fact. It should be viewed as a source of creativity in language, a guarantee of its vitality. Language has always had both the term and the poetic word3. In Lithuanian, hesitation differs subtly from doubt: hesitation emphasizes the inevitable choice between two possibilities, whereas doubt describes a state that encompasses both possibilities. We can say that hesitation is more characteristic of a machine’s mechanical selection according to predefined criteria and marks a branching point in an algorithm. Doubt, on the other hand, describes the method of operation of the instrument of the mind—the human brain. Ambiguity is at the center of attention in contemporary cognitive research (Zeki 2004). This clarifies the place of our research method within the system of the sciences: we are interested in language as a means of understanding the state of the mind. Today, this field is cognitive linguistics in the broadest sense (Cuyckens, Geeraerts 2007). In language, we encounter two different directions of choice. The simplest example is homonymy and synonymy (see Fig. 2). When we hear a polysemous word (for example, baras), we are at a point of doubt. To decide, we usually have to wait for additional information from one context or another. We arrive at a point of doubt in the opposite direction when we have several ways of expressing the same meaning: the bird flies / the bird soars. Again, we decide by drawing on additional information. Depending on the direction, we will distinguish onomasiological ambiguity and semasiological ambiguity4. Homonymy and synonymy are not the direct subject of our research. A polysemous root constitutes semasiological ambiguity. The interaction of semasiological and onomasiological ambiguities lies at the very core of the mechanism of language comprehension. Unambiguity and ambiguity are not two discrete states, but the extreme poles of a dimension that is an important feature of the activity of the human mind. The human mind’s 2 This concept is used very widely. What matters to us is the approach set out by Semir Zeki in his article (Zeki 2004) and elsewhere. 3 Panašiai mano ir Skirmantas Valentas (2012: 12). 4 Cf. Geeraerts 2010: 281–284. 236 Acta Linguistica Lithuanica LXXIV Phonosemantic ambiguity in the Dictionary of the Lithuanian Language FIG. 2. Homonymy and synonymy as semasiological and onomasiological ambiguity The instrument—the brain—is not a computer. The essence of modern computer architecture has not changed since the time of Alan Turing and John von Neumann. It can be said that computer architecture reflects the activity of the conscious mind (Dehaene, Brannon 2011; Zylberberg et al. 2011). At this pole of unambiguity, the human brain seems hopelessly slow compared with modern computers. Yet even the most powerful computers and the most sophisticated programs are still unable to reliably solve even the simplest tasks of the human mind. Von Neumann already surmised why this is so: the human mind does most of its work unconsciously, exploiting parallel analog mechanisms in the brain. Stanislas Dehaene describes the results of the careful research conducted by psychologists of language over the past three decades as follows: “Thus, our unconscious mind is clever enough to store all the semantic associations of a word in memory in parallel and retrieve them from it, even when the word is polysemous or when only one of its meanings fits the context. The unconscious mind suggests; the conscious mind selects” (Dehaene 2014: 66). Through careful contemporary research, psychologists are investigating how children understand words. The renowned psychologist Jean Mandler, summarizing this research, notes that a two-year-old faces a double problem: in which cases to use a word they have learned, and which word to use in a particular case (Mandler 2004: 131). In our terms, the child faces semasiological and onomasiological ambiguity. The word is the same, but to the child it means something different than it does to an adult. In communicating, the child reduces this ambiguity until it reaches the same level as that of the speech community. These and Įžvalgos / Insights 237 ALGIRDAS SAUDARGAS many other studies show that ambiguity is the opposite pole of unambiguity in the human mind, which is immersed in the waters of the subconscious5. As always, Marshall McLuhan (together with the Canadian poet Wilfred Watson) offers a distinctive and fascinating account of the relationship between the conscious and the unconscious in his seldom-remembered book From Cliché to Archetype (McLuhan 2011 (1970)). To illustrate this transition—from an unambiguous, ossified cliché to ambiguity beneath the archetype and then back to a new cliché—he draws on the final lines of the poem “The Circus Animals’ Desertion” by the Irish poet William Butler Yeats: Those masterful images because complete Grew in pure mind but out of what began? A mound of refuse or the sweepings of a street, Old kettles, old bottles, and a broken can, Old iron, old bones, old rags, that raving slut Who keeps the till. Now that my ladder’s gone I must lie down where all the ladders start In the foul rag and bone shop of the heart. He calls that region of our mind, filled with old clichés and their fragments and governed by archetypes that give birth to new clichés, after the poet’s text, rag-and-bone shop of the heart. This was a nineteenth-century English shop selling secondhand goods and waste, trading in rags (rag) and even animal bones (bones). As the editor of the volume explains, even animal bones were a sought-after secondary raw material for the manufacture of fertilizer, glue, or porcelain (McLuhan 2011: vii). That region of the mind is a source of new forms and images—the world of archetypes (McLuhan 2011: 104). Poetic creation boils there; Jacob’s ladder, which Yeats mentions in his poems, rests upon it. McLuhan’s analysis is apt, but the book itself is difficult to read and intricately constructed. What matters to us here is only the unmistakable analogy with what the founders of psychoanalysis called the subconscious and archetypes, although the latter concept is very vague and depends on the author using it. In Lithuanian, we can call the poet’s rag-and-bone shop of the heart the pharmacy of the mind. Dionizas Poška invites “people of discernment” to his Baubliai museum so that they may enjoy themselves, for “the pharmacy of the mind is beneath this roof.” The museum is a source for the imagination, where a craftsperson “of discernment” can find images and ideas for their work. We can play a little with words—after all, váistas and vaistas are words of the same origin, homographs distinguished only by accent. In his Słownik etymologiczny języka litewskiego (hereafter SEJL), Wojciech Smoczyński derives the word váistas from the root *vaid-, shared by Baltic and Slavic, meaning ‘knowledge (about incantations against illnesses)’ (SEJL 2007: 713), while the form vaistas has the secondary root vais- of the verb vysti (išvydo), which he derives 5 Another important topic connected with the dichotomy of unambiguity/ambiguity is brain asymmetry (McGilchrist 2009). 238 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne iš prokalbės šaknies *ueid- (SEJL 2007: 762, 763). Vaistas (be kirčio) yra abeją, apimanti vaizdinę (vaistas) (LKŽ: Dievas kalbėjo Izraeliui vaiste) ir veikliąją (váistas) dalį, kuri tinka nurodyti į archetipinį informacijos pobūdį. Jungas tai vadino lotynišku žodžiu imago – pasąmonės turinių telkinio (komplekso) suvokimo schema. Imago apima įvairius pojūčius (kaip šiuolaikinėje psicholingvistinėje literatūroje vadina – modalumus): tai ir regimo vaizdinio schema, ir garsinė melodija ar ritmas, ir lytėjimo, kvapo ar skonio motyvas, bet taip pat ir ritualinio (schematiško) judesio pajautimas savame kūne. Tai ir vaistas, ir váistas. Proto vaistinė (vaištinė ir váistinė vienu metu) tai ta pati pranokėjų Patacko ir Žarskaus vydija (tos pačios šaknies naujadaras!), tai Freudo ir Jungo pasąmonė, tai XX a. pabaigoje psichologų eksperimentuose tiriama implicitinė atmintis (Mancia 2006), tai McLuhano poeto Yeatso kūryboje surasta „sendaičių krautuvėlė širdyje“, tai tie sąmonei tiesiogiai nepasiekiami atminties klodai, iš kurių James’as Joyce’as du dešimtmečius jam vienam težinomais būdais traukė Finegano budynių į kitas kalbas neišverčiamus sakinius. Daugelis žmogaus dvasios tyrinėtojų mėgina pasidairyti proto vaistinėje. Ten pat žvelgia ir Skirmantas Valentas (1997; 2007; 2012) savo trilogijoje, gilindamasis į Antano Baranausko, Vlado Braziūno, Sigito Gedos ir kitų poetų kūrybinę virtuvę. The bricoleurs “with an opinion” were (and still are) creators of myths. What Claude Lévi-Strauss (1997) writes about mythical thought is part of everyday contemporary interdisciplinary discourse: “Even in our time the bricoleur6 is someone who does everything himself, with his own hands and using, unlike a specialist in a particular field, all sorts of means, often accidental ones. [...] Thus mythical thought is a kind of intellectual bricolage” (Lévi-Strauss 1997: 29). “[...] mythical thought, like bricolage at the practical level, constructs structural wholes, using not other structural wholes directly, but the substratal remains and fragments of events [...] splinters and remnants—as if they were fossils reflecting the history of an individual or society” (Lévi-Strauss 1997: 34). “In a reduced model, unlike when one tries to know an object or being in its real dimensions, the whole is known first and only then the individual parts” (Lévi-Strauss 1997: 37). We have looked around enough in the fields of poetic and mythopoetic creation. We wanted only to make sure that we are far from alone in searching for a “sacred language.” In the pharmacy of the mind, old clichés and their fragments float in the waters of the subconscious in the form of images and fragments of images. These are not only visual images, but also combinations of sounds (words, phrases, or melodies), smells, sensations, tastes, and impressions of various emotions. These images emerge and fade, combining in every possible 6 In the Lithuanian edition, the French word bricolage is translated as “bricolage.” Įžvalgos / Insights 239 ALGIRDAS SAUDARGAS combinations. It is from there that not only poets but also mathematicians draw ideas (Hadamard 1945). Who organizes these images, and what is their grammar? Phonosemantic phenomena are distinctive in that meaning is drawn into unconscious processes. These phenomena have been studied for more than a century. One of the founders of psychoanalysis and creator of analytical psychology, Carl Gustav Jung (with Franz Riklin), began his scholarly career with word-association experiments at the Burghölzli Clinic in Switzerland in the early twentieth century. He continued the experiments of Kraepelin and Aschaffenburg, which established that when rested, subjects associate words by meaning, whereas when tired, they do so by sound similarity. Jung refined the conditions. It became clear that sound-based, that is, phonological, associations predominate when consciousness is subdued and its attention dispersed—for example, just after waking from sleep or simply when bored during an experiment. Jung concluded that focused attention suppresses sound associations, while distraction makes them increasingly numerous, until only sound associations remain (Kugler 1980: 2–7). Paul Kugler (1980: 15) therefore asks in his dissertation: are there autonomous groups of associations in our language, linked phonetically and endowed with meaning in an archetypal image?7? Freud, too, paid considerable attention to sound associations and slips of the tongue. His 1909 account of the Rat Man case is often cited (in which German words are associated: Ratten, Raten, Spielratte, Rita, Heiraten, Rathaus8. 1.3. The Living Mind “Well, of course it is happening inside your head, Harry, but why on earth should that mean it is not real?” (Albus Dumbledore)9 The brain research discussed next also has a thirty-year history. Around the same time that academician Slavenas was drawing his diagrams, in San Diego, California, at the University of California, near the emerging famed Silicon Valley, the Hungarian-born American scientist Marta Kutas wanted to find out how the electrical potential of the human brain (EEG) reflects a word’s semantic relationship to the context of a sentence. Modern technology makes it possible to record the brain’s response to a spoken word, even though a single word 7 A couple of years later, Paul Kugler developed his reflections in the book The Alchemy of Discourse, which was reissued two decades later (Kugler 2002). 8 Freudo aprašytus atvejus būtina vertinti kritiškai (žr. Sulloway 1991: 255). 9 J. K. Rowling. Haris Poteris ir Mirties Relikvijos. Vilnius: Alma littera, 2008, 572. 240 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne The reaction signal is sometimes ten times weaker than the EEG itself. For decades, scientists have been studying such reactions, called event-related potentials (Eng. event related potential – ERP). An ERP lasts up to a second and manages to fluctuate several times in characteristic peaks and valleys. Each maximum or minimum has its own properties, and describing them would take many pages. We are interested in one turn in research that took place in the late 1970s. The aforementioned scientist and her colleagues wanted to find out how the ERP changes, for example, when the last word in the sentence I shaved off my moustache and beard is changed from beard to city (Kutas, Hillyard 1980). The sentence becomes semantically incongruous. The researchers noticed that the second peak changes; from then on, it has been called N400. In the nearly quarter-century since this observation, N400 has become a marker of semantic activity in the brain (Münte et al. 2000). It turned out that an increase in the amplitude of this marker measures the effort required to fit a word to the context of a sentence. For such studies, it is most convenient to present sentences similar to the one mentioned above. The final word may be expected (as in beard) or unexpected (as in city). This test is called a cloze test (Eng. cloze). The key point that immediately became clear is that N400 responds not only to unexpectedness but to a wide range of parameters reflecting the fit between a sentence and a word. After three decades of careful research, a general view roughly like this has emerged. When we hear a word or read it on a screen, we perceive it with fairly precise timing. During the first quarter of a second, the sensory input is analyzed until the word is recognized. In the second quarter of a second, reflected by the N400 marker, the perceived word is fitted to the context relevant at that moment. What follows may be called corrections. Regarding the first quarter of a second, it is important to note that all sensory information received up to that point is used immediately in recognizing the word. Moreover, psychological experiments and neurophysiological measurements have shown that, as a word unfolds, there is continuous, direct access to semantic information, without waiting for the word to be recognized. The second quarter of a second is a precise time window during which the already recognized word is fitted primarily to the sentence context, but also to more general knowledge relevant at that moment. The time window for fitting does not depend on the properties of the information in the word and the context. The time available for perceiving the word in context is given, and once it has elapsed, we will have achieved whatever result we managed to reach in that time. It is also important to remember that everything happens unconsciously. If we fail to make sense of it, we will have to refine our understanding in the light of consciousness. What is this context from the perspective of a neural network? Without a doubt, it consists of the neuronal ensembles of words activated by the sentence heard or seen, as well as other activated neural networks reflecting grammatical, prosodic (related to stress, pitch accent, duration, intonation, or the like), or non-linguistic information, and content retrieved from long-term memory that has triggered associations. The most important point is that the brain actively fits the new word. The brain Įžvalgos / Insights 241 ALGIRDAS SAUDARGAS mechanisms constantly make predictions, trying to anticipate what information awaits them. A word is not mechanically incorporated into, or added to, the existing context. The word falls into a network that is already waiting for it. The example mentioned above merely shows that the word beard had been predicted and expected, and therefore did not even need to be checked. This was reflected in the low amplitude of the N400 marker. The unexpected word city, like any other word that does not fit the sentence context, burdens the brain with the task of finding at least some connection to the context. This is reflected in the maximum N400 amplitude, indicating that the brain used up the entire time window allocated for fitting in its search for meaning. Of course, the context is not always so restrictive as to yield a single, clear prediction. Context is constantly changing, and its constraining properties change and pulse. Likewise, the brain’s predictions pulse from wholly indeterminate to very specific. The brain makes predictions at every level, from visual recognition to understanding language and conversational intentions. The most interesting aspect of language perception is the form taken by the prediction of an expected word. It turned out that the brain’s response is not of just two types—expected word and unexpected word. There is a whole spectrum of intermediate responses between them. The N400 amplitude can be not only maximal or minimal, but can also occupy any intermediate position. In reality, a word heard may be not only expected or unexpected, but also somewhat expected or somewhat surprising. For example, the brain predicts a word category and expects typical members of that category more than atypical ones. If the context points to the word being the name of some kind of fruit, an apple will be more expected than a pear, and even more so than some exotic fruit. The brain expects more common words more than rarer ones, and so on. As far as I know, the N400 marker has not yet been measured with Lithuanian words, but it would be easy to do so and test the reasoning presented below. To bring out the significance of this research for our problem, let us formulate an analogous example in Lithuanian. Consider the following sentence: Every morning the sun rises. The beginning of the sentence, Every morning rises, is highly constraining. If asked to complete the sentence, many people will mention the sun. An unexpected ending to the sentence might be something like this: Every morning an elephant rises. It turns out that the brain expects words similar to the most likely word not only semantically, but phonologically as well. This was demonstrated in research over the past decade by Kara Federmeier, a student of Marta Kutas, together with her doctoral student Sarah Laszlo (Federmeier, Laszlo 2009). I will explain their conclusions using the same Lithuanian example. As they await the most likely cloze completion—sun—the human brain activates not only semantically related words or words that could be associated with it, moon, water, girl (drawing on the polysemy of the verb rises), but also phonologically similar words. These might include words that rhyme with sun: island, resin, soul, saliva or bench. The facilitation of fitting such words is based on the fact that not only semantic-context associations are predicted, but also the phonological context, which need not have any semantic interpretation. 242 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne The likely result of the experiment should be as follows: sun – minimum N400 amplitude; elephant – maximum; a bench or month – medium. This is a fact about the organization of the language system in our brains, demonstrated experimentally. Why do the brain’s responses activate not only semantic but also phonological associations? Much evidence indicates that a word is recognized sequentially, one phoneme after another. The brain works by encompassing broad, simultaneous streams of parallel information. Not only words but also phonemes are predicted until a final choice has been made. Thus, recognizing a word activates the entire cluster of phonologically neighboring words. By the time a word has been recognized, the neural ensemble of a word phonologically similar to the expected one has already been activated. It is thought that this recognition strategy increases reliability and effectively exploits the brain’s advantage—the principle of parallel information processing. We have examined the results of one psychology experiment. This is a field of research called neurolinguistics, or, in a broader context, cognitive neuropsychology. The specialization of modern science has created deep rifts between different disciplines, while the bridges built across the conceptual chasm through careful experiments and close theoretical analysis are unsafe and often collapse irretrievably. Perhaps it is only a matter of time before brain research and artificial intelligence fill all the chasms in the near future, and we feel the firm ground of indisputable science beneath our feet? Will there still be room for the linguistics and psychology familiar to us? In 2016, Brain forum in Lausanne presented brain research programs underway in the United States, Europe, China, and Japan, which had already secured funding in the billions. Some of these programs and other similar research are surveyed in the volume edited by Gary Marcus and Jeremy Freeman, The Future of the Brain (Marcus, Freeman 2015; cf. Adolphs 2015). In the volume’s afterword, Christof Koch and Gary Marcus presented a forecast for neuroscience fifty years into the future (Koch, Marcus 2015). Predicting advances in cognitive science, they argue that five decades from now, thinking, language, social relations, and other phenomena characteristic of the human mind will still not be explained on the basis of neural brain activity. In their view, it will still be unclear how the brain encodes the meaning of a word or the sense of a sentence, despite nanorobots by then being able to penetrate all the way to a living neuron. Thus, for the next half-century, we will have to rely on hanging bridges... Such a sober perspective encourages us not to dismiss any insights, whenever and by whatever means they may have been obtained. Let us compare the conclusions we can draw from an ERP study in Silicon Valley with the psychological experiments conducted a century ago at the Burghölzli Clinic. Measurements of the N400 marker showed that the human brain predicts the end of a sentence (for example, sun), which activates an entire fragment of the semantic and phonological network associated with it (for example, month or bench). This information is compared with the word presented in the experiment. Įžvalgos / Insights 243 ALGIRDAS SAUDARGAS Everything happens unconsciously. Only one final interpretation of the sentence’s meaning is presented to consciousness. Jung and Riklin’s experiments tell us something more. Participants in the experiment are asked to give an association for a specified word. The results show that associations can be both semantic and phonological. This is consistent with the findings of the ERP study. But research conducted at the Burghölzli Clinic showed that the nature of an association depends on the subject’s state in relation to consciousness. Jung drew the important conclusion that the conscious mind seeks semantic associations and suppresses phonological ones. Phonological associations break through when the subject’s consciousness is subdued. Similar results were obtained in studies of patients with amnesia, who could not remember the fact (impaired explicit memory) that they had played a visual computer game (for example, Tetris), but saw images of the game or dreamed about it before falling asleep (Stickgold et al. 2000). Psychologists explain this as the manifestation of implicit memory. Contemporary studies of consciousness distinguish five types of unconscious brain activity (Dehaene 2014: 190–197). According to the “global workspace” hypothesis, a thought that is consciously perceived is widely “broadcast” throughout the brain by large numbers of neurons. Only strong, so-called suprathreshold signals can enter the focus of consciousness. There may be many such signals in the brain at once, but consciousness selects only one coherent sequence at a time. These signals do not disappear; they await the focus of consciousness and can be amplified for global broadcasting. “The unconscious mind proposes, the conscious mind selects” (Dehaene 2014: 66). These strong signals, still awaiting the focus of consciousness, are unconscious brain contents of the first type. The second type consists of signals too weak to attract the focus of consciousness. These are subthreshold signals that disappear and remain inaccessible to consciousness. They cannot attract the focus of consciousness by themselves, but during their brief existence they can affect information that is consciously perceived later. The third type of unconscious brain activity encompasses all the automatic processes of our nervous system, such as breathing. The fourth type encompasses all forms of primary sensory coding, which must be recoded before entering the field of consciousness. We are most interested in the fifth type of unconscious knowledge, which Dehaene calls dormant. This information is encoded in synaptic connections. It begins to accumulate before birth. During the first years of life, although most of the neurons in a human brain are already in place, synaptic connections are formed at great speed. In the first two years, the brain grows to half the size of an adult brain—from one quarter to three quarters of its volume. Yet we remember nothing from the first three years of life (so-called infantile amnesia). Where has this information gone? This is what dormant memory is. The information is still there, but it is inaccessible to conscious memory. There lie the entire grammar and phonology of language, so deeply embedded that the leaders of billion-dollar brain research projects do not expect to reach them in the next fifty years. 244 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne The dormant memory is the same pharmacy of our mind. It is not directly accessible to consciousness, but may be reached by way of intuition. Both poets and mathematicians turn to this source of creativity. The French mathematician Jacques Hadamard (1945), drawing on the experience of one of the greatest mathematicians, Henri Poincaré, which has by now become classical, divides this dialogue into four stages: initiation, incubation, illumination, and verification. A mathematician will not receive inspiration without lengthy, arduous preliminary work, which often fails to produce the desired result. The mathematician works with forms of thought consciously apprehended, acquired through long study, and refined through intense reflection. During the incubation period, combinations of the mathematician’s conscious forms of activity are immersed in an ocean of archetypes, governed by schematically structured instruments lying in the subconscious—dormant, yet real. Illumination (inspiration) comes when the subconscious casts an unexpected solution into the light of consciousness. Its form corresponds to the shape described by psychoanalysis—the archetype. It always carries an emotional charge. In mathematics, emotion manifests itself aesthetically: an unexpected solution is always beautiful. 1.4. Phonosemantic Legends “The almost untouched fields of Lithuanian linguistics—the connection between semantics and the phonetic structure of the root” Audronė Kaukienė (History of the Lithuanian Verb, vol. I) “Oh, cats, little kitty, why didn’t you plough the fallow fields, why didn’t you sow the flax?” Juozas Tumas Vaižgantas (Dictionary of the Lithuanian Language) In her study History of the Lithuanian Verb, Audronė Kaukienė writes: “The connection between structure and semantics has long been observed. The connection between phonomorphological structure and concrete semantics is often emphasized.” She goes on to state: “The almost untouched fields of Lithuanian linguistics are the connection between semantics and the phonetic structure of the root” (Kaukienė 1994: 164, 165). These fields of linguistics (mentioned in the subsection heading) are the territory of science that remains unexplored, into which the pioneers mentioned in the first section are venturing. There is no contradiction with known linguistic facts, but the problem remains unformulated in a way acceptable to science. On the other hand, this territory lies fallow only in research on the Lithuanian language. Great numbers of scholars are working there today, using the very latest technology, although, as folk wisdom warns (along with the directors of billion-dollar brain-research projects), the flax will take a long and painful time to grow after it is sown. This is a frontier territory of science, Įžvalgos / Insights 245 ALGIRDAS SAUDARGAS which it is time for Lithuanian-language researchers to enter as well (they need only overcome the lethargy of Vaižgantas’s cat...). As we have been able to see, so far the only way into this territory is across hanging bridges. Specialists in one field, holding safely to their own shore, will not reach the other. Intellectual risk is necessary. A suspension bridge offers no guarantee of safety. Hypotheses may prove unfounded, but science knows no other path. What matters is that the bridge connecting the two fields be securely fastened at both ends. Let us take a concrete example—the pair of Lithuanian verbs lemti / limti. SEJL gives the Proto-language root *lemH– with a question mark, apparently intended to indicate the meaning ‘to break’ (Pol. łamać) (SEJL 2007: 344). We also find this root in MA, from which we learn that this highly productive root in Baltic and Slavic languages has survived in an encyclopedic-scale monograph on the Indo-European spiritual world only in the form of terrifying nocturnal spirits: Latin lemures and Greek lamia (MA 2006: 411). Of course, at the level of frightening children, they may be compared with our laumės. Comparative linguistics, weighed down by the authority of the classical triad (Sanskrit, Greek, Latin), tends to forget Baltic words, although scholars of Indo-European poetics remember them (cf. West 2007: 384, 386). Let us look at this simple diagram (see Fig. 3), which we have seen hundreds of times in texts on economic and sociological research. FIG. 3. Schema of a part in relation to the whole It would be surrounded by other similar diagrams and strewn with numbers. This is an image schema indicating the relationship between a whole and its part. An image schema (Eng. image schema) is one of the principal instruments of cognitive linguistics10. Here it is important for us only to recall that an image schema is not a picture or diagram (which may look entirely different), but a mental archetype to which the picture merely points (Cuyckens, Geeraerts 2007). In LKŽ, we still find a word for a part of a whole: 10 Terminą „vaizdinys“ vartoju pagal Jakaitienės vadovėlį (Jakaitienė 2010: 45). 246 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne FIG. 4. Conceptual metaphor 1 lãmas 1. ‘a part, piece of something, a fragment’11 I How did you swallow such a big piece of cake so quickly! Ms. Verbs with L/M in the root meaning ‘to break’ seem no longer to be in use (for example, lamėti ‘to break’, cf. Slavic lomati). There are many nouns: lãmas ‘a part, piece of something, a plot of land, parcel’, lamatà ‘violent breaking (during a storm)’, lomà ‘a sunken area of land, valley, piece, lump, plot of land, parcel’. The word luomas takes us into the social domain. LKŽ gives the following meanings: ‘a social group formed on the basis of origin, wealth or occupation’; ‘family status’; ‘kind, origin’; ‘area, parcel, piece’; ‘class of numbers’. The feminine counterpart luoma has, in addition to the first meaning of luomo, other meanings too: ‘happiness, fate’; ‘turn, order’; ‘time limit, deadline’. Everyone knows what the word laimė means. We know its connections with the words lemti, likimas. Here we must pause and point to the second chapter of Algirdas Julius Greimas’s comprehensive study, “Aušrinė and Laima” (Greimas 2005b). It reveals 11 In this article, the elementary objects are lexemes from LKŽ (with only a few examples from DLKŽ). Lemmas are printed in boldface and numbered. I use the terms “lexeme” and “lemma” according to Jakaitienė’s (2005; 2010) definition, but in a narrow sense, as structural objects in LKŽ. Įžvalgos / Insights 247 ALGIRDAS SAUDARGAS FIG. 5. The Cycle of Rėda connections between happiness and fortune, happiness and fate, Laima and Laumė in the broad context of Lithuanian folklore. The semantic content of Lithuanian lexicon and mythology is represented in the most general way in Figure 4. The arrows assign a schematic structure to each domain (frame), or transfer it from one domain to another. Comparative linguistics, being uninterested in the cognitive structure of word meanings, impoverishes the spiritual world of the Indo-Europeans to the level of (in this case) children’s bogeymen. This is just one example showing that comparative linguistics ought to accept at least some responsibility for the aforementioned “spoonful of tar.” Let us look at another example, which formal linguistics automatically rejects. Patackas and Žarskus arrange the transformations into the Cycle of Rėda (see Fig. 5) (Patackas, Žarskus 2013: 10; Patackas 2014: 87). 248 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne FIG. 6. Schema of the verb 2 nókti Nothing prevents us from using words that sound similar (nokti – nykti) or are inversions of one another (migti – gimti) to name relations of transformation. Traditional linguistics rejects any connection between such pairs of words and the science of language. Let us take the pair nokti / nykti. Comparative linguistics reconstructs the following Proto-Language roots: *neh₂k- / *nik- (SEJL 2007: 427, 425). These are different roots and belong to different etymological families. Interest in the matter ends there12. As we have seen, linguists are not particularly interested in the cognitive content of semantics. Nevertheless, it would be worthwhile to investigate it. First of all, it must be noted that the semantic opposition attributed in the Rėda circle to the word pair nokti / nykti is encompassed by the verb itself, nokti: 1 nókti 1. ‘to ripen (of fruit, grain)’ 1 nókti 4. ‘to weaken, diminish’ Let us also recall the verb that is falling out of use: 2 nókti 1. intr. ‘to go, follow behind’ 2 nókti 2. tr. ‘to pursue, chase’ We can represent the structure of this set of meanings as two conceptual metaphors (Cuyckens, Geeraerts 2007). The content of the verb 2 nókti, ‘to go, follow behind; to pursue, chase’, is expressed in the diagram in Figure 6. At point P, the pursuer panoko ‘caught up with’ and pranoko ‘overtook’ the other. The first metaphor {2 nókti → 1 nókti 1. ‘to ripen’} transfers the diagram’s structure from the spatial domain 12 Cf. Karaliūnas 1995: 8. Įžvalgos / Insights 249 ALGIRDAS SAUDARGAS to the domain of nature, where point P marks the peak of ripeness. The same metaphor is also reflected in the verb prieiti: 1 prinókti 1. ‘to become ripe, to come to’ prieiti 1. ‘to approach (something) by walking’ He walked to the end of the bridge. prieiti 13. ‘to become ripe, ripen’ The rye ripened earlier this year. The other metaphor involves a value inversion: {2 nókti → 1 nókti 4. ‘to weaken, diminish’} The point of culmination corresponds not to the positive but to the negative (Nōks nōks i nu-nōks (will die)). The verb nykti helps avoid the awkward inversion. Thus, the full cognitive structure of both verbs is described by two such metaphors: {2 nókti → 1 nókti 1. ‘to ripen’} + {1 nókti 4. ‘to weaken, diminish’ ← 1 nýkti} The first example was merely a mild reproach to the linguists working in the field of comparative linguistics for (apparently) paying little attention to contemporary semantics, but this example raises questions. The structure presented is the cognitive organization of lexical usage as reflected in the LKŽ (at a level that is synchronic across several centuries and the spatial distribution of dialects). In terms of its meanings, it naturally encompasses not only words from a single etymological family but also phonological “neighbors”13. How such structures arise is a question for linguists. The development of conceptual metaphors is the subject of extensive discussion (Fusaroli, Morgagni 2009). Cognitive linguistics recognizes the importance of metaphor in language change. What development the semantic structure of nokti / nykti reflects is a question requiring careful study. The verb 2 nókti, which forms the spatiotemporal basis of the conceptual metaphor, has already fallen out of use, although the event denoted by 2 pranókti remains a productive source of new metaphors (this can be verified by entering the word into the Google search engine). The schema presented should also include a point of transformation, beautifully illustrated by the two folk sayings featuring the verb noko: The rue is ripening, the mint is ripening, the little lily is ripening, my young days are ripening like a green little rue (d.) Gs. The rose in the garden is ripening, the little lily is ripening, my little youth is ripening beside a roguish young man LTR (Mrk). LKŽ and DLKŽ give these sentences under the meaning ‘to lose freshness, to wither’ (DLKŽ gives only the first part of the sentence). It is interesting that Professor Smoczyński notes the ambiguity too, translating this sentence in two ways: “The rue was ripening (=maturing), the mint was ripening, the lily was ripening; like green rue, the days of my youth were withering (=withered).” Translation or substitution of the verbs in both cases reveals and records a particular interpretation. The inimitable beauty of folklore lies in its ambiguity: the same verb nokti occurs throughout, but in the first part of the sentence it leans toward the meaning ‘to mature’, while at the end of the sentence the threat of ‘losing freshness, 13 Plg. Federmeier, Laszlo 2009. 250 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne withering’ emerges. This is a point of transformation, inherent in the essence of wedding customs and the Rasos festival. Unfortunately, linguistics in Lithuania remains indifferent to ambiguity. Word meanings stand in paradigmatic relations to one another. In the sentences The flower is ripening or The apple is ripening, the meaning depends on which meaning of the word noksta we choose: 1 nókti 1 ‘to mature’ or 1 nókti 3 ‘to lose freshness, to wither’. However, in the sentence The rue is ripening, the mint is ripening, the little lily is ripening, my young days are ripening like a green little rue, in the SEJL interpretation the same meanings stand in a syntagmatic relation: first it ripens, then it happens. This syntagm operates not at the level of sentence syntax, but at the level of narrative structure. Let us compare three versions of this sentence: The rue is ripening, the mint is ripening, the little lily is ripening, my young days are ripening like a green little rue = ambiguity in folklore; The rue withered, the mint withered, the little lily withered, my young days withered like a green little rue = LKŽ and DLKŽ interpretation; The rue was maturing, the mint was maturing, the little lily was maturing, my young days were withering like a green little rue = SEJL interpretation. The first sentence, as mentioned, is ambiguous, and this gives folklore its openness to interpretation. The second sentence replaces the word nóko with a synonym suggested by LKŽ. The third sentence is a back-translation from Polish. Formally speaking, the LKŽ interpretation is correct. The dictionary reflects the worldview of the farmer, which took shape over not just one millennium. Language is highly sensitive and precise in relation to important features of the lived environment. The primary meaning of the verb 1 nókti is ‘to mature (of fruit, grain)’. You would not say this of a flower. This is clear from the following sentence: The poppy, while flowering like a rose, develops little holes, opens up, and, when ripe, scatters when shaken by the wind B. The poppy flowers first (like a rose), and then ripens. A blossom does not ripen, but flowers until it withers. Fruit, grain, and seed ripen (scatter when shaken by the wind). But youth both flowers, matures, and withers: Petniūnas did not trust the suitors and was reluctant to take a son-in-law into his household, although his daughter kept ripening and ripening until she over-ripened and lost the charm of youth rš. (LKŽ). This sentence, marked in LKŽ as figurative, that is, as a conceptual metaphor, transfers a schema from plants to human biological nature. Youth can be compared with a lily blossom, a ripening apple, and (unfortunately) an overripe fruit or a withered flower. This point of transformation is beautifully and conceptually precisely expressed in the schema of the Rėda cycle, but we are interested in aspects of its linguistic expression. By formally replacing nóko with vyto in the sentence, we lose the entire poetic essence of folklore. The verb vyto has no ambiguity and therefore does not give the reader (or listener—after all, it is a song!) access to the point of transformation (as cognitive Įžvalgos / Insights 251 ALGIRDAS SAUDARGAS (psycholinguistics). The SEJL interpretation, sacrificing precision14, attempts to construct this approach. The turning point is revealed at the conscious rational level, as the unfolding of discourse (the lily was ripening > my young days were fading). Folklore uses the ambiguity inherent in the verb nokti to open an unconscious approach to the turning point. This is the difference between prose and poetry. This undoubtedly ancient content finds no place in comparative linguistics, even though SEJL, reconstructing the form *neh₂k-, has no doubt about the root’s Indo-European origin. Whether such content, central to the worldview of the farmer, came from the proto-language or was created in the Baltic languages, like postpositional locatives, and is now fading away—these are questions raised by cognitive analysis. DLKŽ shows how this broad and deep semantic field is crumbling. Language inevitably changes. What should be preserved, and what does not matter—this is the essential question. I will not comment on or question the importance of the stability of expression (and spelling). We learn language unconsciously, so the mechanisms of unconscious perception are crucial to its vitality. To know what and how to preserve, it must be studied thoroughly. Now let us consider the fourth version of the sentence, in which the final verb nokti is replaced by the verb nykti. Nóko rūta, nóko meta, nóko lelijėlė, nyko mano jaunos dienos kap žalia rūtėlė. It is presented not to replace the song’s lyrics. There is nothing more to add to what has been said about the ambiguity. The expression of this sentence (nokti / nykti) retains the ambiguity in the first part of the sentence (nóko lelijėlė), but reveals the threat to consciousness in the second part (nyko mano jaunos dienos). In folklore, the turning point lies deeply submerged in the polysemy of the verb nokti. This sentence brings the turning point to the level of consciousness (as does the SEJL interpretation), yet it remains submerged in the subconscious through phonological similarity. This is not some new linguistic fact that contradicts the findings of comparative linguistics. It points to a semantically connected reservoir encompassing both proto-language root forms. The semantic whole (gestalt) in language encompasses both etymological nests, which are, phonologically, at an “apophonic distance.” The example Nokti / nykti motivates the term phonosemantic saga. Saga is a term at the level of narrative. This article is devoted to the methodology for identifying and describing such structures. In this example, the phonosemantic saga encompasses the entire continuum of forms in LKŽ: nokti, nykti, neikti, niūkti, [...] niekas, naktis [...] and so on. The foundation of the semantic gestalt is formed by the aforementioned pair of metaphors and the turning point. From the perspective of Paul Kugler’s “alchemy of discourse,” this is a cluster of phonetically similar words unified by an emotionally strongly motivated network of meanings 14 Dojrzewala is translated in the dictionary as ‘matures, ripens’. As we have seen, LKŽ does not assign these meanings to flowers. I would not venture to judge whether the Polish word dojrzewala can be used not only for fruit but also for flowers. 252 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne of the whole—imago. Incidentally, the threat of the turning point associated with nokti / nykti, which is described in detail in The Book of Transformations, is the same as that discussed in the principal example in the psychoanalyst’s study (the loss of virginity; Kugler 1980: 11). However, the complex of unconscious contents described by psychoanalysts lacks structural organization. Jung’s archetypes of the collective unconscious discern only very general features. Here we need to build a bridge to firmer ground—mathematics. The Chilean-born Swiss psychologist Rafael Núñez (currently working at the University of California, San Diego), analyzing Jung’s correspondence with the renowned physicist Wolfgang Pauli, notes a conceptual correspondence between the archetypes of the collective unconscious and the image schemas used in cognitive linguistics (Núñez 2008: 9). At the end of the millennium, Núñez and one of the founders of cognitive linguistics, George Lakoff, revealed the cognitive structure of mathematical thought (Núñez, Lakoff 1998; Lakoff, Núñez 2001). These are the first steps toward the scientific study of mathematical creativity undertaken by Hadamard more than half a century ago, already mentioned above (see 1.3). Today, scholars working in this direction use methods from cognitive linguistics and psychology, as well as neurophysiological measurements of brain activity (see Núñez, Marghetis 2014; Marghetis, Edwards, Núñez 2014; Dehaene, Brannon 2011; Dehaene 2014). 1.5. Crafting Phonosemantic Sagas “Even in our times, a jack-of-all-trades is someone who does everything themselves, with their own hands and using all sorts of means, often improvised, unlike a master [specialist] in a particular field.” Claude Levi-Strauss To summarize, we can say that a phonosemantic saga is a cluster of phonetically similar words whose meanings’ semantic content in terms of space, time, and force dynamics (ELD) forms a whole, often acquiring an emotional charge. Where should one look for such sagas? The answer: LKŽ. Kugler is interested in manifestations of the unconscious in the human psyche and studies cases in which clusters of phonetically similar words can be discerned in dreams or visions. In illness of the psyche (Freud’s described case of Rat Man) or in slips of the tongue, the state of an individual psyche is manifested. Sagas can be sought directly in language in composed texts or in the very structure of the dictionary. Folklore, as we have seen, conceals cognitive structure in the polysemy of a word—in ambiguity. Syntagmatically, it develops sound patterns through inflectional and derivational paradigms (LKŽ: I am a traveler, a little traveler, who has traveled along the path KlpD20). Texts Įžvalgos / Insights 253 ALGIRDAS SAUDARGAS Folk tradition, it seems, does not create sequences of phonetically similar but unrelated words (cf. nóko lelijėlė / nyko mano jaunos dienos). Such texts would have been noticed and studied. The problem can be illustrated with a short artificial text. Let us compare three stories about crows: I. Three crows are waiting. One flew away, another stayed, and the third is still waiting. II. Three ravens waited in a cage. Time flew by. “It is time,” said one raven, and flew out. The second raven stayed. He held cheese in his beak and thought, “I am hungry; they won’t bend me.” The third raven continued to wait. He was held by longing for his native home. “Nevermore,” he croaked. III. Three ravens perched in a cage. Life flowed by. “It’s time,” said one raven and flew out of the cage. The second raven did not fly. He had cheese in his beak and thought, “I haven’t eaten; they won’t make me.” The third raven lingered. He was drawn by longing for his native home. “Nevermore,” he croaked. The subjects of the first story are three crows, and all the verbs have the same consonantal root structure, L/K, although they are not necessarily related, since they belong to different etymological families (laukti < *leuk; lėkėti < *lek; likti / laikyti < *leiku / *liku). The second story contains other words with the same root structure, L/K (laikyti, lenkti, alkti, laukan, laikas). The crows have been replaced by ravens, which carry many associations in folk tradition; the cheese is associated with the familiar plot of a fable, and the phrase “nevermore” reminds admirers of Edgar Poe’s poetry of the famous poem “The Raven” (Eng. The Raven). The third story no longer contains a single word with L/K-type roots, but its plot has hardly changed. No one creates texts of the first type or speaks this way. The second story develops the semantic core of the verbs in the first story, using roots of the same type (except in the final sentence). The third story retains the cognitive structure of the discourse, although all phonological connections have been erased. Such are the texts of language. But the structure of language has not disappeared. As we have seen, psychological experiments and even precise measurements of the brain’s electrical activity show that words are connected in unconscious mechanisms both semantically and phonetically. In this way, phonosemantic units emerge and influence thought and language. Psychoanalysts notice them in their studies. In language, they are embedded in morphophonological paradigms. But not everything is fixed in grammars. Hence the answer: phonosemantic tales lie in the cognitive structure of the meanings of words in the LKŽ. Phonosemantic tales are not found in texts. A phonosemantic tale has to be staged. We will begin with the phonological part, selecting a pool of words with roots of the same structure. Then, using cognitive 254 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne linguistics techniques, we will construct the structure of the semantic content of the pool’s words in terms of space, time, and dynamics. Let us look at Figure 7, which will help show that the method used in this study does not contradict known linguistic facts but rather complements them. We can view the same column of word forms from a linguistic perspective (on the left) or an extralinguistic one (on the right)15. Let us examine each approach more closely. The word forms are arranged vertically according to the root vowel (including the absence of a vowel—ø). The roots of all the words have the same consonant skeleton (which we will denote K/L, and, distinguishing vocalized from non-vocalized forms: K-L and KL-). On the left, the reconstructed forms of the Proto-Indo-European root indicated here and elsewhere (unless stated otherwise) are taken from Smoczyński’s Słownik etymologiczny języka litewskiego. Thus, the vocalized roots belong to two different etymological families, while the verbs kloti, klėsti and klysti are also unrelated to one another. The verbs kelti / kilti, kalti / kulti, kloti, klėsti, klysti are phonologically similar. It can be argued that their phonological similarity is the same as that of root-vowel alternation (ablaut). Kelti / kilti are related words, so their structural difference is attributed to ablaut. Kelti / kalti are unrelated, and their phonological similarity is not attributed to root-vowel alternation, although the root vowels differ in the same way as in kelti / kalnas. The root-vowel relationship in kelti / kloti is not generally classified as ablaut in Lithuanian, but this kind of consonant cluster occurs in cognate words in other languages (Smoczyński 2006). We can group roots according to their consonantal structure (see Fig. 8). The diagram was compiled by extracting from SEJL all etymological families with the corresponding Lithuanian consonant skeleton. We will call a set of roots extending downward from a node in the direction of the arrows a root cluster (ŠP). For example, root cluster K-L consists of roots with the following skeletons: K-L; K-LB; K-LP16. The numbers indicate the number of etymological families. We found seven such families in root cluster L/K, and as many as nineteen in K/L. The core of our study does not depend on the accuracy of this distribution. A different methodology or possible errors in interpreting the source data do not alter the main conclusion—the abundance of discrete groups of meanings that are not linked to one another. 15 Claude’as Vandeloise’as (Vandeloise 2010: 171) savo erdvinių žodžių genezės sistemoje lingvistinį požiūrį vadina vidiniu (an. internal lexical formation), o ekstralingvistinį – išoriniu (external lexical formation) kalbos sistemos atžvilgiu. Jo sistema remiasi ekstralingvistiniu erdvės sąvokų genezės mechanizmu (plg. skirsnyje 2.1 minimą Herberto Clarko P- erdvės ir L- erdvės skirtį). 16 Such notation is ambiguous—it denotes both the consonant skeleton itself and the root cluster that begins with it. This does not cause confusion, since the context makes clear what is being discussed. Įžvalgos / Insights 255 ALGIRDAS SAUDARGAS FIG. 7. Roots with the consonants K/L This is merely the organization of information—no linguistic content is assigned to the ŠP. However, the ŠP is constructed on the basis not of phonetics but of comparative linguistics (for example, lizdas would belong to ŠP S–D)17. 17In this respect, the construction of the ŠP differs from Paul Slavenas’s scheme, which is based on phonetics (see the first section). 256 Acta Linguistica Lithuanica LXXIV Fonosemantinės sąkmės lietuvių kalbos žodyne FIG. 8. Etymological nests with the consonants K/L In LKŽ we find the following principal meanings of verbs: kélti 1. ‘to take something heavy upward’ 1 kálti 1. ‘to strike with a hammer, drive or pierce by striking’ klóti 1. ‘to spread, put, lay something on a surface’ 1 klė́sti 1. ‘to sweep, sweep about, sweep clean’ klýsti 1. ‘to go along a wrong path, unwittingly stray from the right path’ A direct comparison of these meanings seems arbitrary. From the standpoint of traditional or structural semantics, these meanings provide no basis for linking unrelated verbs. Greimas’s method makes it possible to distinguish linear (kelti, kalti) and nonlinear (kloti) movement (see Greimas 2005a: 64; Nastopka 2010: 144)18. The structuralist method, based on the opposition of oppositions, comes to a halt here. The cognitive structure of space is not revealed to it. The perspective of cognitive linguistics does not examine word meanings within a closed linguistic system, but instead draws on general principles of cognition— 18Strictly speaking, according to Greimas’s analysis, the meaning of the verb kloti belongs not only to the semantic axes of spatiality but also of horizontality (laterality). This is irrelevant to the discussion that follows. Įžvalgos / Insights 257 ALGIRDAS SAUDARGAS ing—that is, in our figure, it approaches the word from the extralinguistic (right-hand) side. This perspective does not contradict the facts of comparative linguistics but rather supplements them. The fact that vocalized roots denote linear movement (kelti—to lift from a surface; kalti—to return to a surface), while roots with the cluster KL- denote nonlinear movement, is an empirical fact. For the linguistic system of Lithuanian, this fact is extrasystemic and therefore seems arbitrary. The cognitive method finds a place for this fact within the extralinguistic system of spatial perception. The left end of Figure 7 is securely anchored on the shore of comparative linguistics. Only a highly experienced specialist can propose making any reliable adjustments there. To build a “suspension bridge,” we must also securely anchor the right-hand (extralinguistic) part. As already mentioned, the methods of structural linguistics no longer work here. We remain in a discrete field of etymological nests that are not semantically connected. Within an etymological nest, a semantic connection is obligatory (unless it has faded in the course of language development), whereas between nests, semantic connections are arbitrary with respect to phonology. The ŠP can easily be projected onto the set of meanings in LKŽ. The elementary objects of our study are lexemes from LKŽ. We question none of the definitions of meaning. An extralinguistic perspective fundamentally changes the relationships among lexemes where roots have a deep and rich ELD structure. This does not mean that the dictionary ought to be rewritten by changing the order of lexemes. ELD rearranges lexemes nonlinearly—it is no longer a fragment of a dictionary, but an episode in a semantic network. The right-hand side of Figure 7 depicts the most general ELD schemas corresponding to different word groups. When lifting an object, we remove it from the surface and place it above that surface. When hammering, we do the reverse—we return the object to the surface. Both movements are linear, that is, one-dimensional. When laying something down, we align two-dimensional surfaces. This reasoning will be meaningful when we identify the common set of these schemas, a certain gestalt with respect to whose structure the meanings of words are related or differ. This set is generated by the basic verbs kelti and laikyti in their basic meanings. The conceptual content of the verb kelti refers to the earth’s pull, or gravity. The experience of gravity evokes a spatial reference frame (AF) with a particular structure and constrains the nature of force dynamics. The essential feature is the division of isotropic three-dimensional space into a two-dimensional surface, which remains isotropic, and a one-dimensional vertical direction. Objects have weight, and we lift them vertically upward, but the weight of objects is the same everywhere on the earth’s surface. The task of the second chapter is to show that the meanings of K/L words in the ŠP are interpreted according to AFK-L, in which the two-dimensional surface is unbounded, and L/K according to AFL-K, in which the two-dimensional surface is bounded and enclosed. All the observations in our study, as noted above, are constrained by the principles of the linguistic system (we select only words of Baltic origin, excluding onomatopoeic words and loanwords), but are not restricted by it. In other words, the study does not contradict established linguistic facts, 258 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne but supplements them. Where does this additional information come from? Lexicology distinguishes its object of study from that of grammar. It studies only nouns and their semantic content, which is invariant with respect to inflection. However, ELD structures are most often expressed in grammatical information. Hierarchical systems that generate spatial concepts are illustrated by prepositions in various languages. Talmy bases his classification of force-dynamic patterns primarily on syntactic structures. The grammatical system of Lithuanian also has analogous means of expression, more or less systematized. In our study, the source of information is the distribution of semantic content among phonologically similar roots (including polysemy). In other words, we seek in the lexicon of words the ELD information that usually resides in grammatical forms of expression. Grammatical information is usually created by combining meaningful morphemes. In our case, ELD information resides in the root itself. The word namas is divided into morphemes—the root nam- and the ending -as. The information conveyed compositionally by the morphemes nam- (the concept of a house) and -as (the nominative case) is summed up in the information conveyed by the word namas. This approach is characteristic of structuralist and generative linguistic traditions. Today it is no longer sufficient. This does not mean that the approach is wrong, but that it is limited. It reveals only one aspect of the phenomenon. The cognitive-linguistic tradition takes a different view. First of all, although lexical information (conveyed by the root) and grammatical information (conveyed by the ending) are fundamentally different in nature, both are meaningful (Talmy 2000: 21; Langacker 2008: 18). Grammar is not a set of rules; it has semantic content. The root (nam-) provides access to encyclopedic memory through a frame to which the particular context points. The ending (-as) provides access to the grammatical paradigm: the network of inflectional or derivational forms in which nam- is a phonological invariant. Roots do not have meanings. Only words have meanings. Each word sense is an invariant of usage. This is how LKŽ is organized. Each lexeme provides a lemma (representing an inflectional paradigm), its description (a semantic invariant of usage), and examples of usage. SEJL presents a sequence of phonological transformations from a reconstructed Indo-European or Baltic root to the root of a word. Strictly speaking, only a lexeme has meaning (as described in LKŽ), whereas the semantic content of a word (apart from relatively rare and uninteresting exceptions) consists of the totality of the lexemes belonging to it. This totality is not merely a numbered list of lexemes; it includes the semantic motivation of the relationship among the contents (descriptions) of the lexemes. Thus, the word sense presented as the primary meaning merely represents the entire motivated system of the word’s meanings. The semantic content of an etymological nest encompasses the motivated system of all the meanings of all related words. Selecting a word’s primary meaning poses many synchronic and diachronic problems for lexicographers (Jakaitienė 2005: 2010). All the more so, the “primary meaning” assigned to a reconstructed root rarely adequately represents the semantic structure of the entire nest. Įžvalgos / Insights 259 ALGIRDAS SAUDARGAS This is especially evident when we consider only the semantic content of an etymological nest of Lithuanian (or Baltic) words, while the “primary meaning” is imposed by the authority of the classical languages19. The rich semantic content of the Lithuanian nest, laid out plainly in LKŽ, is simply ignored. A faulty habit develops of thinking of a root (a word root or one reconstructed for a proto-language) as having a meaning, and that meaning is often manipulated. This view of semantics must be changed. More precise thinking (in terms of the science of semantics) must regard the semantic content of the sequence {lexeme < word < root < reconstructed root} as a successive unfolding of uncertainty. The lexeme expresses the ultimate goal of univocality in conscious language. The reconstructed root has the greatest uncertainty. It is from this perspective that we can speak of the meaningfulness of a consonant pair. This is all the more not a meaning, but a motivated set of all words and all meanings. A phonosemantic tale is precisely the meaning of a consonant pair (ŠP)—its motivated semantic content. We can extend the sequence: {lexeme < word < root < reconstructed root < phonosemantic tale}. The connection between consonant-pair (ŠP) and ELD structures is also indicated by the symmetry and the so-called nokti / nykti laws mentioned by Patackas, as well as by Slavenas’s network of roots. These insights of forerunners were not recognized by linguistics because, as noted, phonological relations are extrasystemic, while semantic relations are extralinguistic. Nevertheless, they can be accommodated, for instance, within Ray Jackendoff’s parallel architecture (Jackendoff 2003). The generative structure of morphology and syntax, based on the phonological system, forms the basic principle for systematizing linguistic facts. The autonomous system of knowledge describing language, whose foundations were laid by Ferdinand de Saussure and whose formal rigor was established by the generation of generativists led by Noam Chomsky, underpinned the aspiration to make linguistics an exact science and created the illusion that language could be completely formalized. In the context of the search for artificial intelligence, the efforts of generative linguistics to formalize language were called the cognitive revolution. The current of cognitive linguistics emerged as a reaction to the unproductive formalism of generative methods. Taking an intermediate position as a kind of mediator, Jackendoff proposed a parallel architecture, noting that language contains more than one distinct, independent generative system. The parallel-architecture approach calls for the study of the interconnections among separate generative structures once they have been described. The main generative structures are phonological, syntactic (including morphology), and semantic. In addition to these, Jackendoff also mentions spatial structure (see Fig. 9). The author himself acknowledges that spatial structure also encompasses the perception of time and patterns of force dynamics (Jackendoff 2003). Spatial, 19 Pgl. MA 2006: 396, 404, kur prokalbės šaknys *h₁enek- ir *bher- indoeuropiečių kalbų spektre apima reikšmes ‘imti’, ‘nešti’ ir ‘duoti’ plačiame freime, bet „pagrindinė reikšmė“ parenkama pagal „autoriteta“. 260 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne FIG. 9. Parallel architecture according to Jackendoff phonosemantic FIG. 10. Phonosemantic tale in parallel architecture Įžvalgos / Insights 261 ALGIRDAS SAUDARGAS The perception of the interaction (dynamics) of time and force, and its expression in language, have been studied with increasing intensity in recent decades. In his Parallel Architecture, Jackendoff proposes describing phonological, syntactic, and semantic generative structures and examining the relations between them. For spatial structure, he indicates a connection only with conceptual structure. This is the distribution of conceptual content across various human senses (modalities). He does not envisage connections between ELD structures and syntax or phonology. The direct connection between ELD and the phonological structure of language is precisely the territory where the aforementioned untouched fields of Lithuanian linguistics lie fallow. Our approach proposes the direct expression of ELD structures in phonology. In other words, this approach postulates that the connection between conceptual structure and phonological expression is accompanied not only by a connection through syntax (including morphology), but also by a connection through ELD. Recalling the meaningfulness of grammar (including syntax) continually emphasized by followers of cognitive linguistics, as well as the predominance of ELD information in grammatical content, we can expand Jackendoff’s triangle of Parallel Architecture into a tetrahedron (see Fig. 10). 2. THE TALE OF SPACE AND TIME 2.1. The Perception and Expression of Space, Time, and Dynamics in Language “Thus, the initial image of space is an a priori intuition, not a concept” Imanuelis Kantas (Kantas 1982: 77) Let us ask what is depicted in Figure 11—a flower in the hand or on the hand? Clearly, the choice of the locative case or the preposition ant depends on how we divide up space as we look at the picture: whether the hand is perceived as a container (a closed surface) or as a support (an open surface). It is no coincidence that this situation has recently attracted particular attention20. It brings together the expression of force dynamics and spatial perception. Both the use of the English prepositions on / in and the prepositional systems of various languages have been studied in detail. In language studies, the context in relation to which the location of objects is described is called a reference frame. Psychology studies how, through the senses, one experiences 20 See Evans, Chilton 2010 and the extensive references therein. 262 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne FIG. 11. A flower in the hand or on the hand? The image is interpreted within one reference frame or another. From a semantic perspective, an onomasiological perspective (how the same experience can be named in different ways) meets a semasiological inquiry here (which situations one or another polysemous preposition applies to). Both aspects are revealed by choosing to divide three-dimensional space with a two-dimensional surface. Space is three-dimensional and isotropic. We can locate something in space only in relation to something else. In mathematics and the exact sciences that use it, objects are located relative to a coordinate system. Languages have their own ways of expressing the relations of the Cartesian coordinate system: up / down; forward / backward; left / right (Landau, Jackendoff 1993). Spatial relations are expressed both by lexical and grammatical means. Reference frames may differ, and languages make their own use of them. Herbert Clark’s analysis, presented as early as the 1970s, sets out an approach that directly asks how we perceive space (Clark 1973). I will briefly review this approach. A child is born into a world where they walk upright on flat ground and feel the effects of gravity everywhere. This shapes the structure of their perception of space, which the author calls perceptual space, or P-space. As the child begins to speak, they use spatial language terms for this experience, forming a system the author calls L-space (linguistic space). Clark examines two hypotheses: the correlation hypothesis and the complexity hypothesis. The correlation hypothesis states that L-space must correspond in its structure to P-space. The complexity hypothesis maintains that a child first acquires simpler terms and then more complex ones (see Clark 1973: 28). Of course, many decades of interdisciplinary studies have passed since Clark’s article mentioned above. I have referred to this article because the two hypotheses it formulates define the general context of this chapter. The literature on the perception of space and its expression in language is vast. A general overview and analysis was presented two decades ago by Stephen C. Levinson (Levinson 1996), and a collection of later works was edited by Evans and Chilton (Evans, Chilton 2010). There one can Įžvalgos / Insights 263 ALGIRDAS SAUDARGAS FIG. 12. The division of three-dimensional space by an infinite plane or a bounded surface also find many references to relevant topics. These can be supplemented with an overview of the perception of space, time, and number in the brain (Dehaene, Brannon 2011), as well as Chilton’s recently published monograph (Chilton 2014). The Cartesian coordinate system introduces one axis for each dimension of three-dimensional space. We are concerned with the very first step in reducing the isotropy of space. We will examine the initial step, in which the three dimensions are divided into a two-dimensional surface and a one-dimensional line. In this initial division, the two-dimensional surface remains isotropic. We have two fundamental, simplest choices—an infinite plane or a closed surface. When three-dimensional space is divided by a plane, the remaining dimension is the vertical direction that intersects the surface. The plane remains isotropic. In the case of a closed sphere, the remaining dimension is the radius extending from the center. The sphere also remains isotropic (see Fig. 12). In the structure of Lithuanian roots, this corresponds to mapping the K/L >< L/K inversion onto the reference frames AFK-L and AFL-K. Structures of spatial perception are most often studied at the syntactic level. Systems of prepositions in various languages are examined. So far, I have not found studies of how the initial division of space is manifested in language. In each case, we can assign one or another kind of division to a reference frame. This division is closely related to the classification of reference frames into egocentric and allocentric, but does not coincide with it. We will take this initial reduction of spatial isotropy as our basis for studying Lithuanian roots. The division of space interacts with the perception of time and force dynamics. In frame AFK-L, time is perceived as one-dimensional or as another kind of movement in space. By contrast, in frame AFL-K, time has its own unique nature, which cannot be reduced to a spatial form; this 264 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne gražiai parodo lietuviški veiksmažodžiai laukti ir likti (plg. Galton 2011). Toks erdvės padalijimas glaudžiai susietas su dvejopų jėgų santykiu. Jei perkeliant vieną daiktą judesys perduodamas kitam, tai gali vykti dviem būdais: judesį valdantis daiktas kito atžvilgiu yra atrama arba talpykla. Tai aišku iš mūsų pradinio pavyzdžio. Bet, pereidami nuo sintaksės lygmens prie šaknies sandaros, lietuviškų šaknų semantikoje atveriame gilesnę erdvės, laiko ir jėgų dinamikos santykį. Neurofiziologiniai erdvės suvokimo tyrimai parodė, kad smegenys taip pat linkusios trimatę erdvę dalyti į paviršių ir vertikalią kryptį.21 What has been said can be summarized in a table: AFK-L AFL-K open closed path time space plane + vertical sphere + radius + point time trajectory in space duration at a point force diffuse, acting over a surface, operating along a line concentrated at a point 2.2. Force dynamics in the polysemy of the verbs kelti and laikyti Pasigilinkime į jėgos dinamiką pagal Leonardo Talmy (Talmy 2000: 409; De Mulder 2007). Kadangi šis, kognityvinėje lingvistikoje jau senokai tapęs klasika, metodas lietuviškuose semantikos tyrimuose dar labai retai sutinkamas, išdėstysiu jo esmę smulkiau. Jėgos dinamika nusako santykį tarp dviejų sąveikaujančių jėgų: agonisto ir antagonisto. Agonistas tai tas objektas, į kurį nukreiptas sakinio dėmesys ir kuris arba pasižymi polinkiu išlaikyti esamą stacionarią būseną, arba pats turi potenciją judėti ar kitaip veikti. Antagonistas veikia priešingai agonistui. Susidaro keturios situacijos. Dvi situacijos yra, kai antagonistas nugali agonistą: JDa – the agonist seeks to maintain a stationary state, but the antagonist sets it in motion; JDd – the agonist has a tendency to move, but the antagonist restrains it. The other two situations are analogous, except that the antagonist fails to overcome the agonist, which maintains its stationary state or motion. They are marked, respectively, JDb and JDc. In his monograph, Talmy (2000: 412) primarily investigates force dynamics as manifested in sentence syntax. If a ball (basketball, soccer ball, or golf ball) is lying 21 Žr. trimatės erdvės neurofiziologinių tyrimų apžvalgą (Jeffery et al. 2012). Įžvalgos / Insights 265 ALGIRDAS SAUDARGAS on the ground, it will stay there unless someone kicks it, picks it up, or otherwise sets it in motion. Suppose the wind is strong enough to move the ball. The situation JDa is then described by the following sentence: With a strong wind blowing, the ball rolls across the field. If we place the ball on a slope, it will tend to roll down unless someone holds it: The ball does not roll down the slope because dense grass is holding it back. This is a case of JDd. In both cases, the ball is the agonist, which is the focus of the statement. The wind and the dense grass are the antagonist, which, by entering the situation, changes (or seeks to change) the agonist’s state. Talmy regards these two schemas as fundamental, bringing action and its cause into language. He symbolizes them with the conjunctions because / since: The ball rolls because / since a strong wind is blowing; The ball does not roll because / since dense grass is holding it back. The other two situations describe the antagonist’s unsuccessful influence, and Talmy symbolizes them with the word despite: The ball lies still despite the strong wind; The ball rolls despite the dense grass. We are interested in cases where force dynamics resides in the verb root itself. In other words, we will examine the force-dynamic content of verb meaning. Turning to examples from Lithuanian lexicon, we can represent them in the form shown in Figure 13. Here we compare the primary meanings of kelti and laikyti and their negations: kelti I.1. ‘to lift something heavy upward’ Lift the sack. Can you lift it off the ground? kelti I.10. ‘to plow, to till’ Oh, tomcat, little tomcat, why didn’t you plow the fallow fields, why didn’t you sow the flax? laikyti 1. ‘to hold so that something does not fall or escape’ A maiden stands by his side, holding a silk kerchief in her hand; You did well, little sister, not to hold the horses back. In this figure, the circle represents the agonist and the crescent the antagonist. The arrow or dot in the center of the circle indicates the agonist’s tendency (to move or remain stationary), while the plus sign in the crescent or circle shows who wins—the antagonist or the agonist. At the bottom, a dashed line depicts the final outcome: movement or remaining stationary. In the sentences from LKŽ, the verb conveying force dynamics is in bold, and the agonist is underlined. JDa and JDd reveal the force-dynamic content of the semantics of the primary meanings kelti I.1. ‘to lift something heavy upward’ and laikyti 1. ‘to hold so that something does not fall or escape’. As mentioned earlier, these are the two fundamental schemas because they point to possible changes in the situation: the antagonist wins, setting the agonist in motion (JDa) or stopping it (JDd). In both cases, the antagonist’s crescent is marked with a plus sign. It is unsurprising that these schemas express the meanings of the primary senses (kelti I.1 and laikyti 1). The other two schemas refer to cases in which the agonist wins and the situation remains unchanged. We can derive schemas JDb and JDc from schemas JDa and JDd by applying a logical transformation—negation. Looking at schemas JDa and JDb, we see that they differ only in the transfer of the plus sign from the crescent (the antagonist) to the circle (the agonist). This is the semantics of negation: the antagonist’s efforts are negated, and the agonist wins (maintains its tendency): did not lift (did not plow) the sack (the fallow fields). Similarly, by transferring 266 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne FIG. 13. Force dynamics with negation a plus sign, we move from JDd to JDc (did not hold, did not keep, did not restrain the dappled horses (silk kerchiefs)). The Talmy schema reveals four possible relations between the agonist and antagonist. We are interested in its connection with the perception of space. I will show that, when the set of four schemas is divided by a vertical line, JDa and JDc, which fall on the left, are perceived in frame AFK-L, while JDb and JDd, which remain on the right, are perceived in frame AFL-K, respectively (see Fig. 14). To bring the perception of space into focus, I replaced the negative sentences in JDb and JDc with uses of the intransitive verbs to rise and to remain. Let us look at the intransitive form of the verb to rise. It describes the state of an object. When we raise an object, it rises. However, the examples given in LKŽ and the definition itself indicate otherwise: to rise 1. ‘to move upward’ The moon rises above the maple trees. Įžvalgos / Insights 267 ALGIRDAS SAUDARGAS FIG. 14. Force dynamics and the division of space The moon rises not because someone raises it. Neither in our everyday worldview nor by drawing on astronomical knowledge can we identify a force that raises it according to schema JDa. Nor does the moon rise itself. The moon simply rises. The sentence expresses the moon’s observed movement relative to the maple tree. In other words, force dynamics are no longer present in the sentence. Nevertheless, the pair of words raise / rise prevents their complete absence. Force dynamics have not disappeared but have been pushed to the periphery of perception—to the background. Pure spatial motion occupies the foreground. The essential distinguishing feature of the semantics of raise / rise is the shifting of JD from the foreground (profile) into the background, leaving space-time in the profile 268 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne as a result. Therefore, we can still provisionally place this sentence in schema JDc. Let us compare this situation with the schemas for the primary meanings of to raise (JDa) and to hold (JDd). The moon rises not because someone raises it, but (according to Talmy’s original interpretation) it rises despite gravity. These are all the LKŽ sentences that describe phenomena literally (without metaphorical transfer): Enormous clouds of dust rose upward. In spring, warm mist rises from your meadows. In the schema, having lost its plus sign, the antagonist’s crescent has, on the one hand, come from JDd (gravity is unable to hold the moon rising in the field of vision). On the other hand, it reflects one of the two forces interacting in JDa (gravity is not the force that raises, but the force that impedes raising). As we have seen, the antagonist’s victorious force no longer participates at all, whereas in the losing schema JDc the antagonist’s force is gravity, which in schema JDa is overcome by the raising force. Thus, the semantics of the verb to rise is vertical movement against the background of gravity. This is the reference frame AFK-L evoked by the verb to raise. It should be noted that replacing the verb to rise with the reflexive form to raise oneself works in some cases, but changes the meaning. The reflexive verb remains in schema JDa, with the agonist identified with the antagonist. Other languages most often use reflexive forms in place of Lithuanian to rise (or analogous verbs). In Lithuanian, such a substitution, when it does not express the cognitive structure, disrupts the deep fabric of our language. Similarly, hold / remain can be discussed. What matters to us is that, once JD recedes into the background, to rise and to remain fall into different reference frames. The verb to rise shifts into the expression of one-dimensional motion, beautifully symbolized by the noun road. Frame AFL-K, with the forces removed, encloses one within an inner, closed space without spatial movement. Only time remains. The poetic example clearly shows that the verb to rise has JD in the background of its meaning. LKŽ contains two sentences illustrating the lexeme to rise 2 ‘to ascend upward’: The dappled falcons rose into the sky V.Myk-Put. A flock of birds rose Kl121. In the poem, the dappled falcons cannot rise upward like mist or simply a flock of birds. Therefore, the following lines from Vincas Mykolaitis-Putinas’s poem “Margi sakalai” (Dappled Falcons) reveal the JD pushed into the background: Guiding the fading ember late Mottled falcons rose into the sky. Having spurned the deceitful dreams of the earth, They spread their wings in the sky. It becomes clear that the mottled falcons do not merely rise, but have severed their ties with the earth, which held them, luring them with deceitful dreams, while they are borne aloft by wings spread wide. Įžvalgos / Insights 269 ALGIRDAS SAUDARGAS FIG. 15. The verb pakelti in different reference frames The claim that, when moving from the verb kelti to laikyti and vice versa, we understand their spatial meaning by replacing one reference frame with another, requires closer analysis. Therefore, we will identify the point of intersection in the perception of the polysemy of the verbs kelti and laikyti (the semasiological perspective), where both verbs function as synonyms, and show that the perception of their overlapping meanings differs only in the reference frame (the onomasiological perspective). The meaning of the popular proverb Many hands lift a great burden need only be explained to primary-school pupils. However, dictionary compilers interpret it differently. LKŽ gives it under the sense pakėlti I.1. ‘to be able to lift something heavy upwards’. Meanwhile, the compilers of DLKŽ assign it to the sense pakelti 9. ‘to be able to hold up’. The latter sense corresponds to the LKŽ sense pakelti I.2. ‘to be able to hold, withstand, endure, manage; bear’. The LKŽ senses are adjacent, so perhaps they differ only slightly? Let us look at Figure 15. On the left side of the figure is Atlas, with the vault of heaven laid upon his shoulders. Looking at the image and hearing someone say Atlas lifted the burden, we would choose the DLKŽ sense, since the mood of the image prompts us to think of a static variant, supported by another sentence in the same entry: I’ll put the sack on your shoulders; can you hold it up? The dynamic drawing on the right leaves no doubt that the verb pakelia in the proverb Many hands lift a great burden is understood according to the verb’s primary sense in LKŽ. 270 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne DLKŽ is a dictionary of contemporary language, and such a shift in meaning would indicate a change in interpretation. However, the picture on the right is even more contemporary, as it comes from teaching materials for primary-school children22. Children find it easier to learn a dynamic interpretation of a proverb. For instance, a fifth-grade history workbook contains the following task: the picture shows people clearing a forest, and one must say what form of agriculture this is associated with and how the picture illustrates our proverb (Many hands make light work)23. Pupils who have read about slash-and-burn agriculture will readily recognize it and explain that it is too difficult for one person to clear an arable field from the forest, but that everyone together can achieve the goal. The proverb contrasts the efforts of one person with the collective work of a group (many hands). The verb make light is secondary in this proverb. Many hands can also do other jobs: roll away a large stone, build a tall castle, or cut down trees in the forest. Folk wisdom chose a simple but vivid task—to lift a heavy load. The verb lift also brought its twofold nature into the proverb. The contexts in which the proverb may be used differ even emotionally, but its instructive content does not change. The meaning of the proverb is invariant, independent of the interpretation of the meaning of the verb make light. Therefore, asking which dictionary is right makes no sense, since both interpretations are valid. This is rarely noticed because the mechanisms determining the difference do not reach consciousness. Our aim is to express in JD schemas the transition from the basic to the overlapping meanings of lift and hold. To describe the cognitive properties of meanings, we must delve deep into the elementary movement of the mind. Let us look at the sentences: The pine trees will withstand storms. The waves raged, foamed, writhed like snakes, and it seemed that the ship would not survive the storm. Which JD schema corresponds to the content of these sentences? Recalling Talmy’s description (The ball lies motionless despite the strong wind), we can assert that this is JDb. The pine trees (the ship) remain in their initial state (unbroken) despite the storm: the pine trees (the ship) withstood (survived) the storms. Let us recall that we derived schema JDb by negating JDa. When the antagonist’s action fails or, like that cat, the antagonist gives up acting, the agonist remains in a stationary state: the fur lies on the ground, the fallow fields are untilled. In this case there is no negation, and the schema undergoes a different transformation. One could use a negation to describe the same situation, but the verb would be different: the storms did not break the pine trees (the ship). To bring the two sentences closer syntactically, we must swap the subject and predicate, that is, use the passive voice: the pine trees 22 Kašubienė 2012: 15 skaidrė. 23 IPU 2014: 7. Įžvalgos / Insights 271 ALGIRDAS SAUDARGAS (the ship) was not broken by the storms. All these uses of the verbs convey the same content: withstand (storms); survive (storms); not be (by storms) broken. There is one essential difference. The pine trees or the ship change from agonist to antagonist. Not being broken by storms or withstanding (surviving) storms differ in how attention is redistributed. We may wonder whether the pine trees themselves are not the cause of the storm (whether they cause storms): The pine trees will cause storms. The pine trees will not cause storms. All these meanings are summarized by a schema in which the change between the agonist (underlined) and the antagonist is most important to us (see Fig. 16). For clarity, if we use a meaning obtained through conceptual metaphor instead of the basic meaning (pakelti VII. 2. ‘to make, cause’), we get a sentence about pine trees that do not cause storms on their own, and can compare it with the sentence we obtained by replacing withstand with survive: The pine trees will not cause storms. The pine trees will cause storms. Formally, the sentences are the logical negation of one another, but their JD schemas coincide. They coincide only in schema type (JDb), however; the schemas are applied differently: the agonist and antagonist switch places. Just as the cat has absolutely no intention of plowing the fallow fields, and they remain as they were, so too do pines have no power in themselves to cause storms, and therefore there is no causal link between pines and the origin of storms. When we interpret the same sentence with the same verb will raise / will not raise, but assign the agonist role not to the storms but to the pines, we get an entirely different meaning, although the schema remains the same. This is not merely a formal reassignment, but a psychological act. The assignment of agonist and antagonist roles is linked to the order in which attention is directed. The agonist is not so much the main actor as the one to whom attention is first directed. The antagonist is introduced into an already known situation, and its effect is assessed in relation to the agonist. In the sentence Pines will not raise storms, we first direct our attention to storms and then bring pines into the field of attention, stating that they do not have a causal relation (will not raise =will not cause) to the occurrence of storms. In the sentence Pines will raise storms, the distribution of attention is the opposite. Attention is first directed to pines, after which it is stated that storms will not manage to break the pines. This arrangement of attention (in Lithuanian) does not depend on word order in the sentence, but is reflected in the syntax. In JDa or JDb with negation, the antagonist is the subject. By reversing the focus of attention, we make the subject the agonist. This is a kind of passive situation, although grammatically it is not a passive form. We can say the same about the verb hold. In this case, a form with a different prefix is more appropriate: Pines will hold back storms. The negation of this sentence is JDb, but the affirmative 272 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne ĮDa Pines will cause storms. ĮDb Pines will not cause storms. Pines will raise storms. Pines will withstand storms. Pines will not hold back storms. ĮDc ĮDd Pines will hold back storms. FIG. 16. Force dynamics of the verbs raise and hold The sentence shifts to the ĮDd schema. This kind of alchemy distills the sought-after unconscious movement of the mind. It is the return of the order in which attention is directed. It transforms the meanings of the primary senses (ĮDa and ĮDd) into a coinciding meaning (ĮDb). What is essential to our study is that in the meanings of the verbs raise and hold, only the ELD schema is transformed, while the conceptual content remains unchanged. From a semasiological perspective, the primary senses are transformed by a single movement of the mind into one of the peripheral senses in the arrangement of LKŽ or DLKŽ. Is the empirical fact that these senses are synonyms arbitrary? This is the central question. Arbitrariness can be refuted by revealing the motivation. The semasiological perspective must show what motivates such an extension of meaning. From an onomasiological perspective, we must answer the question of why the same thing is named by different words. In our case, we have synonyms obtained by extending Įžvalgos / Insights 273 ALGIRDAS SAUDARGAS FIG. 17. The polysemy schema of the verb kelti primary meanings. Semasiological motivation must provide the basis for onomasiological motivation. What gives rise to the need for these peripheral meanings? Let us first consider the process expressed by the verb kelti. To lift an object, we must take it from ground level L1 and, overcoming the force of gravity, raise it to a higher level L2 (see Fig. 17). Grammatical means and sentence context can express movement at various stages. Attention may focus on the beginning of the movement (pakelti, atkelti – Ar pakelsi nuo žemės?), its end (užkelti – Kelk maišą į ratus), or the entire trajectory (perkelti). In any case, the movement described is bounded by two stationary states, as physicists would say. The movement can be extended by moving the load still higher through indefinitely many iterations. The peripheral meaning we are investigating arises when we want to apprehend the balance of forces once the lifting has stopped at an intermediate position. Here we encounter the aforementioned reversal of attention. To lift an object from the ground, we are in the role of the antagonist (Kelk pūrą. Ar pakelsi nuo žemės?). Once we have lifted the load, we must hold it in our hands, and the heavy object becomes the antagonist, seeking to overcome our resistance (Uždėsiu maišą ant pečių, ar pakėlsi?). Why has Lithuanian retained the same verb to name these situations? We cannot answer why—this is an interesting problem for future research. Our study merely shows that Lithuanian has carefully and precisely captured this pivotal point in force dynamics, where the interaction of forces intersects with fundamental shifts in the perception of space and time. To express the JDb state at point L2, we choose AF (see Fig. 18). If we say laiko, we choose AFL/K, and the image in space is static, with only time changing. If we say kelia, we choose AFK/L, and the image is dynamic, tracing a path of movement through space. By choosing AF, we reveal the onomasiological idea 274 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne FIG. 18. The superimposition of the polysemy schemas of the verbs kelti ir laikyti (cf. Fig. 19). This is very important information for understanding the architecture of the generative schema of ELD. Language conceptualizes the fundamental, often recurring features of the observed world. However, language must remain open to a changing world. Word meanings then become open to interpretation. Let us turn to folklore. What meanings does the verb kelti have in these verses: Žemė kėlė žolę, Žirgas kėlė balną, Žolė kėlė rasą, Balnas kėlė raitą, Rasa kėlė pasagėlę, Raitas kėlė kepurėlę Pasagėlė – žirgą. Senai motinėlei. In LKŽ, we find the first two lines of these stanzas: kelti I.2. ‘to hold up, bear some weight (pressing something downward); to carry on oneself, on one’s surface’ Žemė kelia žolę, žolė kelia rasą J.Jabl (Rm); Balos paviršius tik žmogų kelia (su arkliais nevažiuoja). The dictionary gives a two-part description. The first part, ‘to hold up, bear some weight (pressing something downward),’ clearly expresses the JDb balance of forces. The second sentence concerns the surface of a pond frozen over with ice, which supports a person Įžvalgos / Insights 275 ALGIRDAS SAUDARGAS FIG. 19. The semasiological and onomasiological ambiguity of the verbs kelti and laikyti (raises), but not the horse, corresponds perfectly to this description. All the other examples given in the article under kelti I.2. correspond to it as well, except for the fragment of our verse. The second part of the description, ‘to carry on oneself, on one’s own surface’, is clearly intended only for it, since such an explanation is redundant for all the other examples. The first part of the description, ‘to sustain, withstand some weight (when something presses down on it)’, does not fit the first sentence very well. Indeed, it would sound unconvincing to explain the meaning of the verb kelia in the sentence The earth raises the grass as meaning that the earth withstands the weight of the grass pressing down on it. The pressure of dew on the grass would also seem rather strange. Essentially, the second part of the description allows the verb kelia to be replaced with the verb laiko: The earth holds the grass. Since we rejected JDb as unsuitable, what, then, accounts for the verbs’ synonymy? Let us see what the verse looks like if we swap the verbs everywhere: The earth held the grass, The saddle held the rider, The grass held the dew, The rider held the little hat The dew held the little horseshoe, For the old mother. The little horseshoe—the horse (The rider raised the little hat The horse held the saddle, For the old mother) Everywhere except in the final phrase, the verb laikyti is used correctly, either literally or figuratively. The verb laikyti reflects the described image more precisely than kelti 276 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne FIG. 20. The Kanizsa cube being described. However, the image remains static. The verb laikyti is used according to schema]Dd. Yet the entire poetics of the song disappears, because the word laikė evokes a static frame AFL/K in our consciousness. Folk poetry chooses the verb kėlė, and with it a dynamic frame AFK/L, through a schema]Dd that is atypical for it (‘to carry on oneself, on one’s own surface’). Moreover, the verb kėlė brings with it its entire polysemy (with]Da and]Db) in the form of ambiguity (recall the proverb!): Rasa kėlė pasagėlę—understand it as you will, that is, in all ways at once. This is a wonderful example of ambiguity in folk poetry, a true Mona Lisa smile hidden in the sounds of language, analogous to the works of Michelangelo or Vermeer in the visual arts mentioned by Zeki (Zeki 2004:). Let us return to the proverb (Many hands lift a great burden). Hearing the proverb without a specific context, we grasp its lesson effortlessly, without delving into a particular interpretation of the verb pakelia. We choose an interpretation when confronted with a specific situation. Depending on the situation, the proverb evokes three fundamentally different images. This change in the mental image is entirely analogous to familiar visual illusions, for example, the so-called Kanizsa (or Necker) cube (see Fig. 20). If we ask what is depicted in the drawing above, we will get an unequivocal answer: it is a cube. The perception of the image in this drawing is analogous to the proverb Įžvalgos / Insights 277 ALGIRDAS SAUDARGAS without the meaning of a specific interpretation. When we begin to examine the meaning of the specific verb pakelia ‘lifts’, we find ourselves in an ambiguous situation analogous to the one depicted in the lower diagrams with dashed lines. When we ask whether the cube in the picture is depicted as viewed from above or from below, it becomes clear that the upper diagram is ambiguous. These two interpretations are incompatible with one another. The neurophysiological basis of the illusion is that the mechanisms in our brains have no clear criteria for interpreting this dual-aspect image, and the image involuntarily flips. It should be noted that this analysis takes place unconsciously. The Kanizsa cube corresponds to the abstract, instructive meaning of a proverb, while highlighting the dotted lines makes one interpretation or the other easier. Similarly, without changing the meaning of the proverb, the specific meaning of the verb pakelia ‘lifts’ evokes an image characteristic of it. There is no doubt that the brain mechanisms responsible for perceiving visual and linguistic images are analogous. 2.3. The structure of ELD in the polysemy of the verb kelti LKŽ gives 45 meanings of kelti, 16 of kilti, and 33 of laikyti. Including prefixed forms, we find about 300 more meanings. The structure of ELD makes it possible to simplify this diversity substantially by structuring the relations among meanings. In this section, we will discuss the polysemy of the verb kelti. We will take the main meanings of the imperfective aspect as our starting point. The structure of the perfective aspect (with prefixes) essentially repeats that of the imperfective aspect. The semantics of the main meaning of the verb kilti with respect to JD was discussed in section 2.2. LKŽ presents the following grouping of the meanings of kelti: I. ‘to take, push something upward; to put something in a higher place’ II. ‘to elevate, increase, strengthen’ III. ‘to build’ IV. ‘to awaken, rouse, encourage; to stir’ V. ‘to appear, sprout, grow; to swell; to spread’ VI. ‘to transport from one place to another; to string, make’ VII. ‘to prepare, organize something’. First of all, it should be noted that most of these meanings (31) are conceptual metaphors: I. 3, 9, 10; II. 2–8; III. 3–6; IV. 1–6; V. 1–9; VII. 1–2. Of course, a conceptual metaphor maps the structure of ELD onto some other domain (see Figure 4). The metaphor itself is not our subject. We are interested in the expression of the structure of ELD in the root in its pure form. The corresponding LKŽ meanings are classified according to the conceptual component of the domain onto which the metaphor is mapped. To tell the truth, this is a rather hopeless undertaking for a dictionary. For the creators of a future Lithuanian semantic network (if the impetus for creating such a network in Lithuania has already 278 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne can be taken seriously—that is a serious task. Studies of the structure and development of metaphors are extraordinarily abundant (Cuyckens, Geeraerts 2007; Fusaroli, Morgagni 2009). We can convince ourselves that the source of metaphors is inexhaustible by recalling the fifth-graders’ exercise from section 2.2. In interpreting the proverb Many hands make light work, the pupils will in fact have to construct the source of the SPT (source > path > goal) schema as a wild forest, the goal as a ploughed field, and fill the path with examples of various activities drawn from their store of knowledge. The ELD structure remains hidden from consciousness. This is the SPT schema for the meaning of the verb lift, illustrated in Figure 21. By changing disciplines and taking account of the pupils’ level of knowledge, we can generate an unlimited number of such exercises for one and the same proverb. The LKŽ, of course, presents so-called dead metaphors, or metaphors at various levels of conventionalization, often turned into linguistic clichés (Lakoff 1987; Lakoff 1993; Allan 2009). This is a broad field of research, and in this article we delve into it only when strictly necessary for the main objectives set out. Let us consider the other 14 meanings. We have already examined the very first meaning I.1 ‘something that is difficult to take upward’ in terms of JD. The immediate consequence of JD for spacetime is a topological event24. Indeed, in order to lift something, we have to overcome the force of the object’s weight. In the sentences Lift the sheaf. Can you lift it off the ground?, the verbs have both process and event aspect. In the command lift the sheaf, there is not yet any information about whether the sheaf will in fact be lifted. Process aspect does not provide such information. The statement is lifting the sheaf is open to both logical consequences that event aspect necessarily reveals: lifted the sheaf or did not lift the sheaf. This openness, however, is not symmetrical, since the verb to lift unambiguously presupposes the information contained in event aspect, because the opposite result of an unsuccessful lifting effort requires negation with its corresponding morphological markers (the prefix non-). At the same time, the verb to lift unambiguously expresses the spacetime consequences of JD, which are actualized in event aspect: the lifted object is separated from the surface with which it was in contact (Can you lift it off the ground?), and moves upward. If the sheaf has not lost contact with the ground, no judging panel will acknowledge that it has been lifted, unless it has been moved, pulled or shifted. In spacetime, the content of the verb to lift is expressed by three image schemas: verticality, SPT, and JD. Their most general relational features are shown in Figure 21. Lexemes shift attention from one schema to another. I.1 focuses attention on the beginning of motion—the topological event (circled in Figure 21, analogous to the schema of lifting off a surface in Figure 7). This is the source of the SPT schema. Let us move on to the lexeme: I.4. ‘to put or carry something to a higher or another place’ Lift the sack onto the cart. 24 “Topology” is a mathematical term whose use in linguistics is not yet sufficiently clear (Talmy 2000: 25; Petitot, Doursat 1998). Įžvalgos / Insights 279 ALGIRDAS SAUDARGAS FIG. 21. Schemas of the polysemy of the verb to lift Here the goal of the SPT schema is in the foreground. The SPT schema itself is still linked to verticality. The next lexeme once again returns us to the source, while JD takes the form of a causal relation (of a topological nature): I.7. ‘to take down’ Kels nu galvelės rūtų vainikelį, maus nu pirštelio sidabro žiedelį. The lexeme I.8. ‘to unhinge, take off its hinges (a door, window, gate...)’ is merely an application of the same ELD to a specific situation. Here it is worth recalling Kazimieras Būga, who emphasized that the Lithuanian verb kelti, unlike its Slavic counterparts, can be used in process aspect (without a prefix) with the meanings pakelti and nukelti (Būga 1958: 477). Certainly, prefixes distribute attention across the component parts of the SPT (source > path > goal) schema, modifying meaning in their customary way. Process aspect is being pushed out of use. It has a deeper essence and is therefore more difficult to preserve in superficial 280 Acta Linguistica Lithuanica LXXIV Fonosemantinės sankmės lietuvių kalbos žodyne in detail-overloaded language. Here and there it is still alive. Here are four more lexemes: VI.1. ‘to transport, move something to another place’ In autumn we’ll have to move the houses to the kolkhoz too Ktk. VI.2. ‘to transport by boat across the water to the opposite bank’ Carry me across the river in a boat J. VI.3. ‘to tether in another place (livestock in a pasture)’ Father went to move the cows Dgl. VI.4. ‘to make, fit (a gate)’ Don’t raise the gates all the way KlvD92. All of these have the same ELD structure. In the foreground (profile), almost only the ŠTT schema remains, without JD and verticality (pushed into the background). When cows are tethered elsewhere, no one lifts them upward. This is a purely spatiotemporal structure of a topological nature. It is detached from the source domain and linked to the target domain. It may seem that the contexts are entirely outdated and that such schemas are no longer needed by anyone today. Quite the contrary. Contexts change; schemas remain. Today, following such a schema, we move files from one smart device to another. This is not a metaphor. Information has no weight. Like that little cow, we untie it from one storage medium and tie it to another. When source and target are linked to two familiar but opposite states of the human body—lying down and standing upright—we have not only two more lexemes, but also a whole cluster of metaphors. These two lexemes: III.1. tr. ‘to help someone stand up, make someone stand, set upright’ III.2. intr. ‘to get up (after falling or sitting)’ we can also interpret as metaphors, but that is another topic. The other two lexemes: I.5. ‘to give someone a higher position, cause to rise’ II.1. ‘to make something higher, more elevated’ occupy an intermediate position. It is worth recalling that I.2. ‘to withstand, bear some burden (when something presses downward)’ I.2. ‘to carry on oneself, on one’s surface’ The earth raises the grass, the grass raises the dew we have already discussed in Section 2.2. One reflexive form remains: I.6. refl. ‘to climb, rise upward’ Mist rose from the meadows. Here we need to recall the verb kilti, whose JD we have already discussed in Section 2.2: kìlti 1. ‘to rise upward’ The moon rises above the wych elms; In spring warm mist rises from your meadows. Teks dar kartą pažvelgti į Būgos raštus (Būga 1959: 113). Jo žodžiai skamba kiek senoviškai, bet jų prasmė labai šiuolaikiška: „Svetimoji kalbos dvasia dau-gio inteligentų smageninėje nustelbė ne tiktai tą dalyvį, bet ir ištisą savaiminų veiksmažodžių (neutra, media) rūšį. Vietoje sakyti kyla, pakilo; [...] daug kas da-bar ima sakyti ir rašyti keliasi, pasikėlė; [...]“. Be abejo, lema I.6 refl. LKŽ yra visai teisėtai. Galima sakyti abe-jopai: Mist rose from the meadows; Mist rose (itself) from the meadows. Įžvalgos / Insights 281 ALGIRDAS SAUDARGAS However, their cognitive content differs. As we saw in Section 2.2, the verb kilo corresponds to the]Dc schema, in which gravity remains in the background and only spatial movement is visible (the verticality schema). The reflexive form, on the other hand, can be interpreted as a kind of variant of the]Da schema, in which the agonist coincides with the antagonist. Compare the sentences: A little cloud rises from the sea, and my hay will lie down. A little cloud rises from the sea, and my hay will lie down. A little cloud rises (itself) from the sea, and my hay will lie down. The first sentence is taken from a song and given under LKŽ kilti 1. In it, we encounter the second of the present-tense forms provided by LKŽ: kȳla, kilsta, kȳlna. In the second sentence, the first form is used for comparison (kȳla), while in the third sentence the reflexive form is used, following the example from the lemma: kelti 6. refl. ‘to climb, rise upward’ A black little cloud rises, and the hay will lie down in the meadow. The cloud with kȳla is passive for the reasons discussed in Section 2.2. The reflexive form activates the cloud and therefore adds vividness (the reflexive]D form comes into play). The verb form kilsta given in LKŽ is homophonous with this lemma: kilst interj. [to mark a sudden upward movement] What is used in the song: the ideophone kilst or the shortened form kilti 1, kilsta? Whatever the case, something in our “brains” resonates with very sensitive and profound things. Perhaps it is time to name them in truly rigorous (not formal!) scientific language? 2.4. The structure of root clusters K-L and KL- I am a traveler, a little traveler, who has journeyed along the road KlpD20 (LKŽ) The silk of leaves covers the earth S.Nėr (LKŽ) A little maple grew tallest, spread its leaves widest (LKŽ) Let us look at the meanings of the verb kalti. Once again, we have a cluster of metaphors: 1 kalti 4, 5, 8, 11–13. The following lexemes remain: 1 kalti 1. trans., intrans. ‘to strike with a hammer, drive by striking, thrust’ 1 kalti 2. trans., intrans. ‘to beat, thrash, thump’ 1 kalti 6. trans., intrans. ‘to chop, hew’ 1 kalti 7. trans. ‘to sharpen, beat (millstones)’ 1 kalti 10. trans., intrans. ‘to hammer metal’ 1 kalti 14. trans. ‘to beat, peen (a scythe), sharpen’ 1 kalti 3. trans., intrans. ‘to clatter, click, clack; to drip’ 282 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne FIG. 22. Schemas of the verbs raise and hammer We see an essential difference from the polysemy of the verb raise. All the meanings have the same ELD structure, while their differences are conceptual (with or without metaphor). What interests us is only the relation between the ELD structures of the verbs raise and hammer (see Fig. 22). The spatiotemporal and JD properties of the verb hammer are suitable for the structure of the AFK/L frame and even fill it. Overcoming the force of gravity, after lifting an object from a lower place L1 and placing it in a higher place L2, we have several options (see Fig. 17). In Section 2.2, we examined the intersection between a static state (time) and further lifting (path) (cf. Fig. 18). In state L2, the object can be released from one’s hands or dropped to the ground. However, while maintaining contact with and control over the object, we can return it to the two-dimensional surface from which it was lifted in only one natural way—by hammering. Įžvalgos / Insights 283 ALGIRDAS SAUDARGAS L-K(T) fly L-(J)K stay, hold, time L-(V)K field I,II L-(V)K wait L-(N)K bend, lean LK elbow LK starve, hunger I,II K-L raise, rise K-L hammer K-L imprison (S)K-LB speak K-L path K-L raised K-L bone K-LP loop K-L neck KL to lay KL to wander, to ramble I,II KL-JD to rave, to babble KL-V to rely on, to hinder KL to ask, to listen KL-B to stick KL-VB to kneel KL-N clan KL-V maple KLŠ kulšė AFL/K 3D space AFK/L Fig. 23. Perception of space in the K/L <> L/K root cluster SEJL places the verb kalėti in a separate etymological nest. Looking at this cluster of LKŽ lexemes from an extralinguistic and extrasystemic perspective, we should have no difficulty interpreting them from the ELD perspective in parallel with kalti: 3 kálti, kãli (kãla), kãlė [see 1 kalėti 3] A large tree crushes a small sapling, or overshadows it. 1 kalėti, kãli, -ėjo 1. intr. ‘to be in prison’ 1 kalėti 3. tr. ‘to smother, suppress, overshadow (plants)’ 284 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne 1 kalėti 4. intr. ‘to waste away while staying in one place, while being pinned down’ Perhaps others, having sensed the heavens, will no longer languish in solitude B.Sruog. refl.: I languish all day, shut up, and grow sad]. 2 kàiti, kãlia, kõlė tr. ‘to put one thing against another, conceal, support’. Let us now look at five nouns that SEJL assigns to separate etymological nests (see Fig. 23). To describe them, I will use the term “reification” approximately in Ronald Langacker’s sense (Langacker 2008: 95, 107, 119), so that our approach remains independent of word formation. To clarify, let us take an example. In tracing the polysemy of the verb kelti, we used the SOURCE > PATH > GOAL schema. In the extended form of the schema, the verb kelti focuses on the beginning and the end, ignoring the process itself. The verb kilti, as we have seen, expresses the VERTICALITY schema. The noun kelias, then, is the reification of a part of the one-dimensional AF frameK/L, wholly independently of the word’s formation or etymology. With respect to the verb kelti, kelias is a reification of the PATH, the middle part of the SOURCE–PATH–GOAL schema. In turn, this schema serves as a prototype for the noun kelias in the form of a conceptual metaphor. The basic usage is concrete: 1 kelias 1. ‘a strip of land along which one walks or travels’ This road is no longer used by people J.]abl. In other uses of the same lexeme, kelias is understood differently: That road runs through Svieriai. This is fictive motion (Talmy 2000: 99). Beside the road, a little green oak grew. Here kelias is part of the landscape. 1 kelias 3. ‘a line in space (air, water) along which movement takes place’ I will guide her along the paths of the stars Mair. This is a conceptual metaphor. The LKŽ description specifies the prototype. This list could be continued, but the schema would be the same throughout. The same can be said of the other etymological nests, with a few comments. Kaulas is clearly a one-dimensional object, although its use exploits the mechanical properties of bone. Kilpa is a one-dimensional closed line. We use Kaklą not only to nod along the vertical axis, but can also look around in the horizontal plane. Interestingly, comparative linguistics comes to our aid here: kaklas < *kal-klas (SEJL 2007: 245). This can be understood in various ways, but according to the AF frameK/L, it is just as it should be. What remains is the archaic word kėlta (this is the form given in SEJL). In LKŽ we find: 1 kėlta 1. see 1 keltas 1 ferry ‘a wooden floating bridge used to cross to the other side of a river where there is no bridge; a raft’ 2 kėlta see keltava 1 kėltava 1. see keltuva 1 kėltuva ‘cattle, livestock’. Įžvalgos / Insights 285 ALGIRDAS SAUDARGAS How does this word fit AFK/L? We have already encountered one keltuvėlė: 1 keltuvė ‘cow’. Let us recall: kélti VI.3. ‘to move (livestock in a pasture) to another place’ The father went to move the cows. This now nearly forgotten activity was, from the time of the Neolithic Revolution, the most characteristic feature of grazing livestock—that is, keltuvai. It is clear: keltuvos are those that must be constantly moved (of course: kélti VI.3). What does comparative linguistics tell us? SEJL derives the word kelias from the form *kelas and the extinct verb *kelti (meaning ‘to set in motion, to cause a push or a turn’), which are linked to the Proto-language root *kʷelh₁-o- (SEJL 2007: 271). The connecting link in this explanation is the word kélta. Thus, according to (SEJL 2007: 272), kelias is ‘a path along which livestock are driven to pasture’. Somehow everything is upside down. But after all, we are looking from different perspectives... In summary, we can assert that the structure of the ŠP K-L ELD everywhere adheres to the restriction of spatial division to a one-dimensional segment. The diversity of ELD is concentrated in the polysemy of the main verb kelti. Further changes in ELD structure link unrelated words (kelti / kalti), while within an etymological nest, ELD variation accompanies word formation and inflection in the usual way. Let us turn to ŠP KL– and consider the remaining part of spatial division—the two-dimensional surface. Understandably, the verb kloti attracts attention: klóti 1. trans. ‘to spread, put, lay something on a surface’ The earth is laid with the silk of leaves S.Nėr. The other meanings have the same ELD structure. It may be worth mentioning just two characteristic metaphors: 22. refl. ‘to go well’ My journey went well. išklóti 7. fig. ‘to recount in full, set out, explain’ One must find such a person, who would know how... to stand face to face and lay everything out properly LzP. like laying leaves ‘to speak fluently’. However, we have ten etymological nests (see Fig. 23). What links them to the two-dimensional surface? We are examining empirical material, so we need not impose anything on it, but we must clarify how it can be interpreted within the accepted theoretical framework. Let us recall that the verb kelti, while remaining within the bounds of one-dimensional segmentation, developed its polysemy within the abstract ŠTT schema, moving entirely away from both vertical motion and JD itself (still the same little cow, or modern-day information). We find the same thing in ŠP KL-. Language developed the opposition {one-dimensional line <> two-dimensional surface} into an opposition between abstractly one-dimensional purposeful motion (ŠTT) and its opposite. Kloti is merely a directly spatial interpretation. In fact, we find two essential generalizations that motivate the entire cluster of KL- vocabulary. The first generalization is the extension of motion from a two-dimensional plane into three-dimensional space (klėsti), with corresponding metaphors (klėstėti). The second generalization is the negation of purposeful motion (klysti): 286 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne to err 1. intr., tr. ‘to go by a wrong route, inadvertently to stray from the right path’. One could spend hours and hours with LKŽ and see that, from the ELD perspective, all the richly proliferating bundle lexemes, both in phonological expression and in nuances of meaning, are motivated in one way or another by the aforementioned means. Below I give only a few examples. 1 to spread, klėsčia (–čia, klėčia, klėčia), -tė (klėtė, klėšo) 1. tr. ‘to sweep, to sweep about, to klastyti’ 3. tr. ‘to rustle, to envelop’ 4. tr. ‘to hide, to protect, to defend’ 6. intr. ‘to move, to flutter’ and so on. No comment is needed. Klesčia (1 klėsti 9. ‘to spread, to expand’), thus both expression and meaning spread and develop through the nuances of a variety of contexts. Among the usages we find: The little maple grew tallest, spread its leaves widest. From the ELD perspective, there is no longer any need to ask what a maple is. klejóti 1. ‘to rave, to babble’ klejóti 3. ‘to walk while lost, to wander, to be unable to find one’s way’ SEJL chose klejóti as its headword. Phonological changes have their own criteria. From the perspective of ELD structure, klejóti 1 is a metaphor whose prototype is klejóti 3. Thus, its typical meaning is the aforementioned klýsti 1. The contrast between mountain and valley is echoed by speaking / listening. This list could be continued. I was unable to find an interpretation only for the pig, which belongs to the K–L cluster. I did not include it in the list because SEJL regards the etymology of kiaulė as obscure. We could claim that the correspondence is one hundred percent, unless some pig gets in the way. But even a pig dreams of clouds (the vertical dimension!). 3. THE NARNIA TALE The heroes of Clive Staples Lewis’s children’s books repeatedly enter the mysterious land of Narnia and return from it after performing various feats. In each adventure, the children get there in different ways, but they all share one feature: Narnia is entered by diving. They do not always dive into water. Once they have to pass (dive through) the back panel of an old wardrobe; another time, they must cross some other real (including water) or imaginary surface. Lewis’s heroes always enter not only another space but another spacetime. Often, they return home not through the same gates by which they entered, but by other means, the twists of the plot interweaving the worlds they have visited with their own into a single worldview. Įžvalgos / Insights 287 ALGIRDAS SAUDARGAS FIG. 24. Root cluster N-R It is said that the writer chose the name of the land of fairy tales by chance. While leafing through an old map of Italy, he noticed the town of Narnia (now Narni) in Umbria, halfway between Rome and Assisi. He liked the sound of the name. It is a coincidence, but Narnia’s connection with the Lithuanian verb nerti is not arbitrary. The Umbrian river Nera flows through the town of Narni and empties into the Tiber, while the root ner means ‘warrior’ in Umbrian. In Lithuania, too, the river Neris flows, and instead of a warrior, beneath the water one dives, nirà or neróvė. This is Indo-European heritage. The Greek sea god is Nereus, and his daughters are the nymphs Nereids. In Scandinavian mythology, the god Njord and his consort Nertus are associated with water. Let us not forget Narcissus’s relationship with the water’s surface that reflects his face (MNM 1992: 212, 231, 201). It should no longer surprise us that we find this cognitive structure only in Lithuanian, since the verb nerti has preserved its primary meaning 1 nérti 1. ‘to plunge suddenly, to descend (e.g. into water)’ only in Baltic and Slavic languages (MA 2006: 293). The phonological consonant schema is shown in Figure 24. In SEJL we find eight entries, which are grouped into four etymological nests: N-R nerti I, nerti II, narys, niuras N-R norėti, norint N-RŠ neršti N-RT nertėti 288 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne From a semantic perspective, etymology groups the N-R sound-symbolic tales into five prototypical meanings: 1 nėrti 1. ‘to plunge, descend suddenly (usually into water)’ 2 nėrti 1. ‘to make lace, a net; to knit’ norėti 1. ‘to feel a desire, to wish’ 1 neršti ‘to release roe, spawn (usually of fish or frogs)’ nertėti 1. ‘to be very angry, rage, fume, seethe’ LKŽ lists the verb nerti in two separate entries: 1 nerti ‘to plunge suddenly into water’ and 2 nerti ‘to knit’. Linguists disagree on whether these are different meanings of a single word or homonyms, that is, words whose meanings are so far apart that they must be presented in separate entries. In his textbook, Jonas Palionis (1999: 185) recalls that the first edition of DLKŽ (1954) adopted the former solution, whereas subsequent editions, like LKŽ, adopted the latter. The textbook explains that “to this day, no stricter criteria have been established to determine whether, in a particular case, we have one polysemous (multivalent) word with meanings that have diverged, or two words—homonyms” (ibid.). In other words, it remains unclear what is meant by the notion of “divergent (or similar) word meanings.” Comparative linguistics does not explain much more in this particular case. SEJL gives the following etymology: 1 to dive I, 2 to dive, niuras, narys < *nerH- norėti < noras < *nārā- < *nā-ra < *neh₂-ro- neršti < *nerḱ- SEJL states that 2 nerti is “probably etymologically related” to 1 nerti, but offers no arguments (SEJL 2007: 421, 422). MA presents 1 nerti as an example of a verbal form preserved in Lithuanian that corresponds to the Proto-Indo-European preposition *ner (after). This preposition gave rise, in an interesting way, to the Germanic word north—north (because the Indo-Europeans associated the north with below). By contrast, 2 nerti is presented in the same work as an example of the root *(s)ner- (meaning ‘to fasten with thread or a rope’) (MA 2006: 234, 293). In fact, the problem is the same: the reconstructed roots have meanings similar to those of the Lithuanian words, so we must address the same question of semantic similarity or distance. The spatiotemporal structure of diving into water (1 nerti) is depicted in Figure 25. For knitting (2 nerti), such a figure is no longer sufficient. In what ways are the spatiotemporal structures of the two groups of meanings similar, and in what ways do they differ? Both are shown together in Figure 26. Let us imagine a surface P stretched across a closed line L. The movement begins at point A and crosses surface P at point B. Next, point C is marked as the point at which one chooses between two possibilities for continuing the movement. If surface P is the surface of a body of water and line L is its shoreline, the movement can continue beneath the surface; but in order to return to starting point A, we will have to cross the (water) Įžvalgos / Insights 289 ALGIRDAS SAUDARGAS FIG. 25. The spatiotemporal structure of the verb 1 nerti at point D on the surface. If P is a (conceivable) surface on a closed line, we can turn in the other direction and, having gone around boundary L without crossing the surface, return to starting point A. It is precisely at point C that the choice between 1 nerti and 2 nerti takes place. That point of choice is the object of our study. What kind of object is it, and what sciences study such objects? The drawing is readily interpreted in mathematics. It depicts an important property of a two-dimensional bounded surface: having passed through such a surface, we have two paths by which to return to the starting point. This situation can be formulated as a fact of a branch of mathematics—topology. However, mathematicians are entirely uninterested in the fact that these two possibilities are expressed by the same Lithuanian root. Nor does semantic research require mathematics. The spatial relations are self-evident, and the complex field of mathematics will add nothing here. Everyone sees the same meanings of the verb, which are described in detail in LKŽ, but sees only what the paradigm of a particular science is able and wants to see. The meanings of the verb nerti clearly fall into two groups of concepts that are fundamentally different: 1 nerti (to plunge suddenly into water) and 2 nerti (to knit). Diving into water has nothing to do with knitting. What does this approach fail to see? It fails to see what is the principal methodological tool of our study: a component of the content of a word’s meaning, 290 Acta Linguistica Lithuanica LXXIV Fonosemantinės sąkmės lietuvių kalbos žodyne FIG. 26. The combined spatiotemporal structure of the verbs 1 nerti and 2 nerti which expresses the relations of space, time, and force dynamics (ELD). Let us place the winding cat at point C so that it can see the whole ELD structure. The object of our study is not so much the point itself as the state of mind when faced with a choice. In Section 1.2, we called such a state abēja. Looking at our example in this way, we not only see the full range of the word’s meanings, but can also discern a distinctive fact about the Lithuanian language. The root ner has retained its state of abēja in Lithuanian. This can readily be verified by unfolding a sentence such as the following: I am knitting… I am knitting a net… I am knitting a net into the water. One could use the straightforward psychological tests performed by the thousands every day around the world to check how the understanding of the sentence changes. It is very likely that it would change in the following sequence: 1 nérti > 2 nérti > 1 nérti. Until psychologists in Lithuania become interested in the unique properties of their own language, we have no choice but to continue the research using our own means. At point C, we choose a spatial context, Įžvalgos / Insights 291 ALGIRDAS SAUDARGAS called a reference frame (AF). The choice prompted by this and other Lithuanian examples is undeservedly ignored in contemporary research on spatial cognition. Let us recall that space is three-dimensional and isotropic. The semantics of the verb nerti lead to the same first step for reducing spatial isotropy that we discussed in Section 2.1. The two simplest possible choices distinguish the meanings 1 nérti and 2 nérti. The meaning 1 nérti is understood in the AFK-L frame, since the surface of the water through which one dives is not bounded in any usage. Unlike K/L semantics, where everything takes place above the ground surface, here the state described is beneath the (water) surface. The meaning 2 nérti is no longer associated with water, but the surface through which one passes is bounded. We move not into the AFL-K frame, as in the case of L/K, but into the intersection of the AFK-L and AFL-K frames. Here the term “intersection” should be placed in quotation marks, since it directly describes only the geometric fact that contour L is formed as the cross-section of an open and a closed surface. For the combination of the reference frames themselves, one should perhaps use the technique of conceptual blending (Eng. blending). This would take up too much space in this article. 4. THE TALE OF LOVE AND HAPPINESS The structure of the L-M/M-L root cluster is presented in Figure 27. The headwords in SEJL are grouped into 11 etymological nests: L-M to determine, happiness M-L dear, affectionate, to love M-L to grind M-L to grind, mill, woodpecker, a kind of rye, flour, clay, blue, a mixture of grain and chaff, sand M-L lie M-LD to pray M-LD a prayer M-LK firewood M-LK firewood M-LŠ to scrape M-LŽ to milk, giant The L-M root cluster comprises only one etymological nest, whose semantics, expressing the part–whole relationship, we reviewed in section 1.4. Here we will examine in greater depth the M-L root cluster, which consists of ten etymological nests. I will review the semantics of only the four main nests (without additional consonants). The tale’s title hints at symmetry love / happiness. Yet the cognitive nature of symmetry lies not in these popular words, but in the reversal of the relationship between the whole and the part. Reversing the ELD structure lemti / limti—the breaking off of a part (lamo) from the whole (lėmimą)—we should expect the opposite action: forming a whole from parts. 292 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne FIG. 27. Root cluster L-M / M-L The M-L cognitive structure says something different. As we shall see, the semantics of M-L words leads to the relationship between bounded / unbounded and discrete / continuous. The verbs in the M-L cluster express a transformation from the discrete bounded to the discrete unbounded, followed by a transformation into the continuous bounded. Indeed, when breaking a piece off a cake or loaf of bread, we do not remove a component from a mechanism, but separate a portion of a bounded continuous object (break off). It would be interesting to examine and reconsider Greimas’s mythological analysis in this light. In this article, I will limit myself to an analysis of the M-L phonosemantic cluster, to which I now turn. The semantics of M-L / L-M takes the form of a conceptual metaphor (as understood in cognitive linguistics). It is a mapping from a concrete domain (bread making and consumption) to an abstract one (community relations), linked by the same ELD structure. Bread making (and consumption) consists of four stages: a) grains are ground into flour, b) flour is mixed (with water), c) bread is baked (in a fire), d) bread is sliced for consumption. These stages are not invented, but are taken from the language. They correspond to three etymological nests, at the root of which are the verbs: malti, milti and lemti. All the other nests are based either on nouns (for example, mylas), or contain consonant clusters (for example, melsti), and are derivational from both cognitive and phonological perspectives. These Įžvalgos / Insights 293 ALGIRDAS SAUDARGAS The three stages in the domain of community relations are not obvious, so I will first present ELD. Cognitive semantics divides perceived things into bounded / unbounded and discrete / continuous (Talmy 2000: 59). The four resulting combinations are shown in Fig. 28. The arrows indicate a sequence of transformations: a shift from bounded discrete to unbounded discrete, a transformation from discrete to continuous, and, finally, the transformation of unbounded continuous into bounded continuous. Our aim is to ascertain that the ELD structure of the semantics of verbs with the root M-L corresponds precisely to these transformations. Moreover, these transformations occur in parallel both in the concrete domain (breadmaking) and in the abstract domain (community relations), as required by the conceptual metaphor schema. At first glance, it is not obvious that the verb malti is associated with the transformation of bounded discrete into unbounded discrete: malti 1. ‘to grind grain into flour’. This meaning clearly indicates a simple mechanism by which flour is obtained from grain. Its use in a broader context, ‘to grind, rub, crush something into flour, into small particles’, is even clearer. Malti means turning larger material objects into smaller, finer ones or (in the case of flour) even barely perceptible ones. From this perspective, the verb malti is synonymous with the verbs grūsti, trinti, smulkinti, skaldyti and the like, which express the same process in various respects and degrees. Let us look more broadly at the context provided by LKŽ. That miltai means ‘a product of ground grain’ tells us nothing new either. But the singular form of this word changes a great deal: miltas ‘the floury part of a grain’. Since this is crucial to our study, it is worth looking at a set of examples provided by LKŽ: Oats have little flour, but a lot of bran. When a grain is bitten through, the flour can be seen. Whether a bean or a pea, it has the same kind of flour. The seed of small flax is so full of it, and has a large floury part. The grains dried up, but were not yet in the flour, hence the famine [1868]. These examples present the structure of a grain, which was very clearly understood in the farming community. It follows that we obtain miltus not by “grinding grain into flour”, but in some other way, because miltas is already inside the grain. It was even said that “the grain is / is not in the flour”, depending on whether the miltas is large or small. Grains already contain miltus within themselves. Every grain has a miltas, and the whole of them, when ground, forms miltus. Miltas is inside the grain as in a container called sėlena. When ground, miltas is separated from the sėlenos and released from its container. Let us use cognitive structures to clarify how the meanings of the verb malti ‘to grind grain into flour’ and ‘to grind, rub, crush something into flour, into small particles’ differ, although LKŽ presents them as the same meaning 294 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne to grind: bounded discrete becomes unbounded discrete 2 to flour: unbounded discrete becomes unbounded continuous 1 to flour: continuous unbounded turns into continuous bounded FIG. 28. Change in the M/L structure of the tale variants. We can crush stones into crushed stone; finer still is gravel, then sand, and finally dust. By conceiving of the grinding of grain in this way, we are at a coarser level of cognitive structure, one that suits stone but overlooks the deeper insight into the structure of a grain. If we grind stone, the continuous bounded becomes discrete unbounded. More precisely, this is another schema, in which a singular object becomes plural. If we grind grains and attend to the structure of a grain, the discrete bounded becomes discrete unbounded. These transformations are understood differently. A stone (continuous bounded), ground into dust (discrete unbounded), is not a container for that dust. A grain has bran, which contains flour; by grinding grain, we release the flour from that container. These are fundamentally different cognitive structures. Language expresses this very precisely. When releasing discrete or continuous contents from a container, we use different words. We pour water (continuous) from a vessel, but pour grain (discrete) from a sack. We cannot sprinkle water or pour out grain. Flour, like dust, is neither poured nor poured. It occupies an intermediate, transitional position, so Įžvalgos / Insights 295 ALGIRDAS SAUDARGAS we have to use a universal word: we pour flour (but we also pour both water and grain). Thus the cognitive structure of the semantics of grind contains both a coarser and a deeper structure. Whether grind is synonymous with pour or crush depends on the depth of insight. Language can be both precise and poetically free. As we have noted, we cannot pour water, nor pour grain. Transferring the same schema metaphorically into the social domain, we find in LKŽ that people can pour, stream, and pour themselves. In this way, we substantiate the previously mentioned stage a of bread-making. The structure of stages b and c is entirely evident, since bread is made from flour (discrete unbounded), which, through kneading and baking, becomes a continuous object. Incidentally, the word bread is used both as bounded (a loaf of bread) and unbounded (give me some bread). This stage comprises two transformations: from discrete to continuous and from unbounded to bounded. We will consider them together because the semantics of the words overlap in an interesting way. The connection with the verb milti is not obvious. LKŽ gives two meanings (and two phonological forms): 1 milti, milia, mylė ‘to put heated stones into cold water (in a bathhouse)’ Let’s go to the bathhouse—I saw that he’s already putting the stones in. 2 milti, milsta, milo ‘to begin to love’ Living in harmony, people begin to love one another. The second meaning, 2 milti, reveals the cognitive layer. The meaning given in LKŽ, ‘to begin to love’, remains on the surface, whereas the example points to a deeper layer. People milsta if they live in harmony. We can understand this as people beginning to love one another. But the verb milsta is used intransitively, that is, it reveals not what people do, but what happens to them—people milsta. As they mildami, people move from an assemblage of separate individuals (discrete unbounded) to a community (continuous bounded). This is the same transformation as when bread is made from flour. LKŽ does not give the meaning that flour, as it turns into bread, milsta. Again, it does not matter at all whether such a meaning existed at some point or had some other phonological expression. What is conservative is the cognitive structure itself, and the meaning of the verb bears witness to it. Most remarkable is that the first meaning of the word, 1 milti, is directly connected with bread-making. Flour no longer has individuality, but without it you cannot bake bread. Yet, to become bread, flour must undergo a procedure of fire and water, rich in symbolism. The mystery of fire and water begins, and only through it will we have fragrant bread. Thus, flour is not merely the result of grinding, but the cause of a subsequent process (milti). They fuse the constituent elements into a single whole—bread. It remains only to recall that in ancient times bread was baked in the bathhouse... The cognitive structure of stages b and c is also attested by other words. SEJL gives the word mylas as the headword for all meanings and forms associated with love, including 2 milti. Only 1 milti is listed separately, with its origin left unexplained. Meanwhile, the cognitive perspective puts everything into place. Nouns cluster around: mylas, mielas, mielės. It is not difficult to discern the common denominator of meaning—the gathering together into 296 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne a single whole (bread, community, building) as a “ferment”—would that mean yeast, plaster, or a dear person? This is a transformation from the discrete to the continuous. The final stage d is associated with the phonological transformation M-L > L-M. As mentioned, we discussed the L-M pair in section 1.4. Interestingly, LKŽ also contains the phonological metathesis milinys / liminys. Understandably, its nature is entirely different. ŠALTINIAI DLKŽ – Dabartinės lietuvių kalbos žodynas, trečiasis elektroninis leid., vyr. red. Stasys Keinys. Vilnius: Lietuvių kalbos institutas. IPU 2014–2015 – Istorija. Pilietinis ugdymas. Mokymosi priemonių katalogas. Kaunas: Šviesa, Vilnius: Alma littera. LKŽe – Lietuvių kalbos žodynas 1–20, 1941–2002; elektroninis variantas. Redaktorių kolegija: Gertrūda Naktinienė (vyr. redaktorė), Jonas Paulauskas, Rიტutė Petrokienė, Vytautas Vitkauskas, Jolanta Zabarskaitė. Vilnius: Lietuvių kalbos institutas. Prieiga internete www.lkz.lt. MA 2006 – Mallory J. P., Adams D. Q. The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World. Oxford University Press. MNM 1997 – Mify Narodov Mira. Olimp. SEJL 2007 – Smoczyński Wojciech. Słownik etymologiczny języka litewskiego, Vilnius: Vilniaus universitetas. LITERATŪRA Adolphs Ralph 2015: The unsolved problems of neuroscience. – Trends Cogn Sci 19(4), 173–175. Allan Kathryn 2009: An Inquest into Metaphor Death: Exploring the Loss of Literal Senses of Conceptual Metaphors. – Conceptual Metaphor Theory: Thirty Years After. Journal of Cognitive Semiotics 5(1–2), ed. by Riccardo Fusaroli, Simone Morgagni, 291–311. Būga Kazimieras 1958–1962: Rinktiniai raštai 1–3. Vilnius: Valstybinė politinės ir mokslinės literatūros leidykla. Chilton Paul 2014: Language, space and mind: the conceptual geometry of linguistic meaning. Cambridge: Cambridge University Press. Clark Herbert 1973: Space, time, semantics, and the child. – Cognitive development and the acquisition of language, ed. by T. Moore. New York: Academic Press, 27–63. Įžvalgos / Insights 297 ALGIRDAS SAUDARGAS Cuyckens Hubert, Geeraerts Dirk 2007: Handbook of Cognitive Linguistics. Oxford University Press. Dehaene Stanislas 2014: Consciousness and the Brain: Deciphering How the Brain Codes Our Thoughts. Viking Press. Dehaene Stanislas, Brannon Elizabeth, ed. 2011: Space, Time and Number in the Brain. Searching for the Foundations of Mathematical Thought. Academic Press. De Mulder Walter 2007: Force dynamics. – Cuyckens, Geeraerts Handbook of Cognitive Linguistics. Oxford University Press, 294–317. Ellenberger Henri F. 1970: The Discovery of the Unconscious: The History and Evolution of Dynamic Psychiatry. New York: Basic Books. Evans Vyvyan, Chilton Paul, ed. 2010: Language, Cognition and Space: The State of the Art and New Directions. Equinox Publishing Ltd. Federmeier Kara, Laszlo Sarah 2009: A beautiful day in the neighborhood: An event-related potential study of lexical relationships and prediction in context. – Journal of Memory and Language 61, 326–338. Fusaroli Riccardo, Morgagni Simone, ed. 2009: Conceptual Metaphor Theory: Thirty Years After. Journal of Cognitive Semiotics 5(1–2). Galton Antony 2011: Time flies but space does not: Limits to the spatialisation of time. – Journal of Pragmatics 43(3), 695–703. Geeraerts Dirk 2010: Theories of Lexical Semantics. New York: Oxford University Press Inc. Greimas Algirdas Julius 2005a: Struktūrinė semantika. Metodo ieškojimas. Vilnius: Baltos lankos. Greimas Algirdas Julius 2005b: Lietuvių mitologijos studijos. Vilnius: Baltos lankos. Hadamard Jacques 1945: The Psychology of Invention in the Mathematical Field. New York: Princeton University Press. Jackendoff Ray 2003: Foundations of Language: Brain, Meaning, Grammar, Evolution. USA: Oxford University Press. Jakaitienė Evalda 2005: Leksikografija. Vilnius: Mokslo ir enciklopedijų leidybos institutas. Jakaitienė Evalda 2010: Leksikologija: studijų knyga. Vilnius: Vilniaus universiteto leidykla. 298 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne Jeffery et al. 2012: Navigating in a 3D world. – Behavioral and Brain Sciences. Cambridge University Press. Kantas Imanuelis 1982: Grynojo proto kritika. Vilnius: Mintis. Karaliūnas Simas 1995: Etnokultūra, kalba ir krikščionybė. – Liaudies kultūra 3, 8–10. Kaukienė Audronė 1994: Lietuvių kalbos veiksmažodžio istorija 1. Klaipėda: Klaipėdos universitetas. Kašubienė Laima 2012: Marijampolės „Smalsučio“ pradinė mokykla. Lietuvių kalba. Skaitymas. Literatūros bei kultūros pažinimo pradmenys. Patarlės ir priežodžiai. 3 klasė. Prieiga internete: http://peliukai.lt/lietuviu-kalba/3-klase Koch Christof, Marcus Gary 2015: Neuroscience in 2064: a look at the last century. – The Future of the Brain, ed. by Marcus G., Freeman J. Princeton University Press, 255–269. Kugler Paul 1980: Archetypal linguistics: an inquiry into the relation between phonetics and the imagination. Thesis/dissertation. State University of New York at Buffalo. Kugler Paul 2002 (1982): The Alchemy of Discourse: Image, Sound and Psyche. Daimon. Kutas Marta, Hillyard Steven 1980: Reading senseless sentences: Brain potentials reflect semantic incongruity. – Science: 207, 203–205. Lakoff George 1987: The Death of Dead Metaphor. – Metaphor and Symbolic Activity 2(2). Lakoff George 1993: The Contemporary Theory of Metaphor. – Metaphor and Thought, ed. by Andrew Ortony. Cambridge: At the University Press. Lakoff George, Johnson Mark 1999: Philosophy In the Flesh: The Embodied Mind and Its Challenge to Western Thought. New York: Basic Books. Lakoff George, Nuñez Rafael E. 2001: Where Mathematics Comes From: How the Embodied Mind Brings Mathematics into Being. New York: Basic Books. Landau Barbara, Jackendoff Ray 1993: “What” and “Where” in Spatial Language and Spatial Cognition. – Behavioral and Brain Sciences 16, 217–38. Langacker Ronald 2008: Cognitive Grammar. A Basic Introduction. Oxford University Press. Levinson Stephen C. 1996: Language and space. – Annu. Rev. Anthropol. 25, 353–82. Lévi-Strauss Claude 1997 (1962): Laukinis mąstymas. Vilnius: Baltos lankos. Įžvalgos / Insights 299 ALGIRDAS SAUDARGAS Maceina Antanas 1994 (1978): Filosofijos kilmė ir prasmė. Raštai 6. Vilnius: Mintis. Mancia Mauro 2006: Implicit memory and early unrepressed unconscious: Their role in the therapeutic process (How the neurosciences can contribute to psychoanalysis). – Int J Psychoanal 87, 83–103. Mandler Jean 2004: The foundations of mind. Origins of conceptual thought. Oxford university press. Marcus Gary, Freeman Jeremy 2015: The Future of the Brain: Essays by the world’s leading neuroscientists. Princeton University Press. Marghetis T., Edwards L., Núñez Rafael R. 2014: More than mere handwaving: Gesture and embodiment in expert mathematical proof. – Emerging Perspectives on Gesture and Embodiment in Mathematics, ed. by L. Edwards, F. Ferrara, D. Moore-Russo. Charlotte, NC: IAP-Information Age Publishing. McGilchrist Iain 2009: B001KE101Y Master and His Emissary: The Divided Brain and the Making of the Western World. Yale University Press. McLuhan Marshall, Watson Wilfred 2011 (1970): From Cliche to Archetype. Gingko Press. Münte Thomas F., Urbach Thomas P., Düzel Emrah, Kutas Marta 2000: Event-related brain potentials in the study of human cognition and neuropsychology. – Handbook of Neuropsychology 1, ed. by F. Boller, J. Grafman, G. Rizzolatti. Nastopka Kęstutis 2010: Literatūros semiotika. Vilnius: Baltos lankos. Núñez Rafael 2008: What is Mathematics? Pauli, Jung, and Contemporary Cognitive Science. – Recasting Reality: Wolfgang Pauli’s Philosophical Ideas and Contemporary Science, ed. by H. Altman-spacher, H. Primas. New York: Springer. Núñez Rafael, Lakoff G. 1998: What did Weierstrass really define? The cognitive structure of natural and epsilon-delta continuity. – Mathematical Cognition 4(2), 85–101. Núñez Rafael, Marghetis K. 2014: Cognitive Linguistics and the concept(s) of Number. – The Oxford Handbook of Numerical Cognition, ed. by R. Cohen Kadosh, A. Dowker. Palionis Jonas 1999: Kalbos mokslo pradmenys. Vilnius: Jandrija. Patackas Algirdas 1983 (1979): Saulė ir kryžius. Kaunas: savilaida. Patackas Algirdas ir kt. 1984: Ethos. Baltų kultūra ir lietuviškoji savimonė. Kaunas: savilaida. 300 Acta Linguistica Lithuanica LXXIV Fonosemantinės sakmės lietuvių kalbos žodyne Patackas Algirdas 2014: LITUA. Lituanistika, publicistika, beletristika. Vilnius: Versmė. Patackas Algirdas ir kt. 2014: Lietuva yra vydija. Kaunas. Patackas Algirdas, Žarskus Aleksandras 2013 (2002): Virsmų knyga. Kaunas. Petitot Jean, Doursat René 1998: Modèles dynamiques et linguistique cognitive: vers une sémantique morphologique active. – Rapport du CREA 9809, Ecole Polytechnique. Rзzauskas Dainius 1997: Lingvistinė beprotybė ir šventoji kalba. – Liaudies kultūra 1, 10–18. Rзzauskas Dainius 2006: „Etnosofai“A. Patackas ir A. Žarskus. – Liaudies kultūra 2, 45–47. Saudargas Algirdas 1986: Žodžio duris pravėrus. Kaunas: savilaida. Slavėnas Paulius 1984: Gamtamokslinis žvilgsnis į prokalbes. – Mokslas ir gyvenimas 7, 34–35. Slavėnas Paulius 2001 (1985): Prokalbės prasmingumas. Akademikas Paulius Slavėnas. Vilnius: Vilniaus universitetas, 259–278. Smoczyński Wojciech 2006: Laringalų teorija ir lietuvių kalba. Vilnius: Lietuvių kalbos institutas. Stickgold Robert, Malia April, Maguire Denise, Roddenberry David, O’Connor Margaret 2000: Replaying the Game: Hypnagogic Images in Normals and Amnesics. – Science 290, 350–353. Sulloway Frank J. 1991: Reassessing Freud‘s Case Histories. The social construction of psychoanalysis. – ISIS 82, 245–275. Talmy Leonard 2000: Toward a Cognitive Semantics 1. Concept structuring systems. Cambridge, MA: MIT Press. Valentas Skirmantas 1997: Lingvistinis pasaulis poezijoje. Vilnius: Mokslo ir enciklopedijų leidybos institutas. Valentas Skirmantas 2007: Mė(lynojo)nulio lingvistika Vlado Braziūno ir Sigito Gedos poezijoje. Vilnius: Baltos lankos. Valentas Skirmantas 2012: Kalbos stigmos poezijoje: (prō)nōmen. Šiauliai: Šiaulių universiteto leidykla. Vandeloise Claude 2010: Genesis of spatial terms. – Language, Cognition and Space, ed. by Evans Vyvyan, Chilton Paul. Equinox Publishing Ltd., 171–192. Įžvalgos / Insights 301 ALGIRDAS SAUDARGAS West M.L. 2007: Indo-European Poetry and Myth. Oxford University Press. Zeki Semir 2004: The neurology of ambiguity. Consciousness and Cognition 13(1), March, 173–196. Zylberberg Ariel, Dehaene Stanislas, Roelfsema Pieter R, Sigman Mariano 2011: The human Turing machine: a neural framework for mental programs. – Trends Cogn Sci 15(7), July, 293—300. Fonosemantic Sagas in the Dictionary of the Lithuanian Language SUMMARY Several decades ago fonosemantic phenomena have been noticed in the Lithuanian language. It was observed that the senses of certain words with interchanged consonants remain related. Those observations were not supported by the scholars of the Lithuanian language. In this article an account explaining fonosemantic phenomena in the Lithuanian language is presented. This account partitions the content of the word sense into the conceptual part and the content which expresses space, time and force dynamic (STD) aspects. The conceptual part of the word sense provides access to encyclopaedic knowledge and is processed consciously. STD part remains unconscious. The fonosemantic phenomena under this account are explained as the extension of the cognitive topology of the network of word senses to phonologically similar though not genetically related words. Several examples are investigated. The network of senses of the Lithuanian verb kelti (to rise, to lift) is studied and expanded. It is demonstrated that STD content of verbs with consonants K/L in the root corresponds to the frame of reference which partitions isotropic 3D space into unbounded 2D manifold plus verticality of one dimension. On the other hand, STD content of senses of the verb laikyti (to keep, to hold) corresponds to the frame of reference which partitions isotropic 3D space by bounded 2D manifold. Words with consonants N/R; M/L <> L/M are analysed in a similar manner. Įteikta 2016 m. gegužės 5 d. ALGIRDAS SAUDARGAS Europos parlamento nario Algirdo Saudargo biuras Teatro g. 8, LT-03107 Vilnius, Lietuva [email protected] 302 Acta Linguistica Lithuanica LXXIV