The LAnguAge of LinguisTic reseArch : is There rooM for MeAning exTension ?

Meaning extension, including metaphors and metonymies, has traditionally been associated with the language of fiction. Cognitive linguistics has given this area a new impetus by unravelling the metaphoricity of our everyday language and various discourses ranging from spoken to written, from learner language to academic or professional (legal, medical, political). Lakoff (1991 and 1995), Cienki (2005), Schmidt (2003), Cibulskienė (2006), Vaičenonienė (2002) and many other researchers seem to give preference to political and socio-economic discourse which has generated within this branch of linguistics a large amount of research all over the world. Academic discourse (AD) and its multiple layers of meaning, however, has received less attention since it is generally believed that AD aims at discovering the truth, explaining and arguing, it is rigid, critical and unambiguous. According to Hyland (2004, 87), AD


Introduction
Meaning extension, including metaphors and metonymies, has traditionally been associated with the language of fiction.Cognitive linguistics has given this area a new impetus by unravelling the metaphoricity of our everyday language and various discourses ranging from spoken to written, from learner language to academic or professional (legal, medical, political).Lakoff (1991 and1995), Cienki (2005), Schmidt (2003), Cibulskienė (2006), Vaičenonienė (2002) and many other researchers seem to give preference to political and socio-economic discourse which has generated within this branch of linguistics a large amount of research all over the world.
Academic discourse (AD) and its multiple layers of meaning, however, has received less attention since it is generally believed that AD aims at discovering the truth, explaining and arguing, it is rigid, critical and unambiguous.According to Hyland (2004, 87), AD depends upon the demonstration of absolute truth, empirical evidence or flawless logic.Its persuasive potency is seen as grounded in rationality and based on exacting methodologies, dispassionate observation, and informed reflection.(…) We see this form of persuasion as a guarantee of reliable knowledge, and we invest it with cultural authority, free of the cynicism with which we view the partisan rhetoric of politics and commerce.
So this paper attempts to provide evidence for the claim that AD, like any other discourse, is equally inclined to meaning extension, specifically, metaphors and metonymies.The paper is structured around several key issues: first, it gives background to the study, including a discussion on academic discourse, its relation to other discourse types, the contrastive parameter of investigation and other methodological issues and second, it identifies major types of metonymical and metaphorical extensions as reflected in the data corpus and manifested in linguistic texts in English (EN) and Lithuanian (LT).

Academic discourse: between fiction and non-fiction
Discourse is understood in this paper as linguistic and non-linguistic (social) event with the functions of linguistic/language items rendered and realisable only in social contexts.This understanding is in line with Schiffrin et al's approach (Schiffrin et al 2001), which emphasises social features; discourse could also be treated beyond the scope of language (e.g. when dealing with the discourse of racism or the discourse of power (ibid.)).
AD, as claimed by many other authors, is confined to its own, academic, discourse community and has a clearly identifiable addressee.AD circulates within and is meant for the AD community.Traditionally, AD has given preference to formal register; it aims at one meaning and maximum disambiguation of the text.
Differently from AD, fiction is addressed to 'any and all', irrespective of beliefs, values, professional background or affiliation etc.In terms of register, fiction is generally informal or neutral; sometimes it might equally embrace all registers.Another feature of the fictional text is its well-expressed and often intentional polysemy, or multiplicity of meaning, at the word, collocation, sentence or text level.Fictional texts are often referred to as creative and imaginary, which is why they can vary in structure, length and vocabulary and why their language is polysemous to a very large extent.Many new meanings that evolve in the text are metaphorical in nature.Interestingly, Steen (2004) has pointed out that the more fictional the text, the more alert the reader is to its metaphoricity.
So if we assume that fiction and non-fiction make up two extremes of a continuum, academic texts could logically be put close to the non-fiction end of it.Reasons for that would include such features as rigid structure, critical attitude, aimed mainly at explaining, arguing and thus contributing to the accumulation of knowledge.AD also serves as a medium of communication among scholars.At the semantic level, the overall aspiration of AD is one and only understanding of the same phenomenon by all members of the community.Thus, the language should be unambiguous and monosemous.This would equally apply to subject-specific terms and probably to a smaller extent, to the general academic language (semi-technical and technical in Garret's terminology; Garret quoted in Ellis 2006, 436).

AD and other discourses
It has been generally accepted in Cognitive Linguistics that not only fictional texts, but also socio-economic and especially political texts are extremely metaphorical.Economic discourse seems to be generated/understood in terms of sickness (health) (see Boers 1999, Urbonaitė andŠeškauskienė 2007), and political discourse is mainly seen as a battle or sports competition (Lakoff 1991and 1995, Cibulskienė 2006).In socio-economic discourses language seems to be revealing: the surface covers but is indicative of well-hidden underlying principles or true intentions, which sometimes turn out to be the opposite from what is expressed on the surface.
Interestingly, Hyland (2004) defines AD in terms of rationality, dispassionate observation and other features which would indirectly point out that AD is much less prone to metaphoricity than any other discourse.On the other hand, Tannen (2002) claims that AD is a battle, adversative and agonistic.She admits that the battle is ritualised, which implies that participants of AD tend to compete among themselves but follow certain rules of politeness.Thus, such an approach would not exclude metaphoricity.

Linguistic AD and cross-linguistic aspect of research
In a vast range of disciplines, linguistic AD is specific in that its object and instrument of research partially overlap.It aims at investigating natural language phenomena using the same natural language as an instrument, or metalanguage.So the use of natural language in the second function is bound to create ambiguity and multiple meanings eventually disambiguated by the context.
The cross-linguistic and cross-disciplinary methodology of research within the Conceptual Theory of Metaphor has generated a considerable amount of research.It has helped identify interesting cross-linguistic and language-specific peculiarities in different discourse types (cf. Schmidt 2003, Cibulskienė 2006, Urbonaitė and Šeškauskienė 2007 etc.).The CTM has proven an efficient tool for translation (see Schäffner 2004, Marcinkevičienė 2006).

Data and methods
The present pilot research has been limited to a fairly small corpus of data amounting to ca. 20 000 words in EN and ca.18 000 in LT.The EN data was collected from the journal of Applied Linguistics (2006); the LT data was drawn from the journal Kalbotyra (2006).The sources are representative since they both are acknowledged research journals, articles there are published after careful editing and reviewing.
The main methods of research have included hypothetical deduction, contrastive analysis and the Conceptual Theory of Metaphor (CTM).The CTM is based on the assumption that metaphors underlie surface (linguistic) metaphorical expressions.Interestingly, numerous expressions might refer to one and the same metaphor, or one metaphor can generate numerous linguistic expressions, sometimes very language/culture specific.So generally, metaphors are mappings between source and target domains, in other words, between two spheres of human experience (Lakoff & Johnson 1980/2003, Kövecses 2002, Ungerer and Schmid 1996 etc.), for example, life (target) is seen as a journey (source) in the following example: (1) He had reached a crossroads in his life.(BNC)1 Crossroads in the above case refer to difficult situations in one's life (=journey), which in another language could be easily conceptualised in a completely different way.

Terminology
Meaning extension presupposes any meaning beyond the primary meaning of a word, collocation, sentence or a string of text/discourse arising either from co-text or context or existing in a system.Two major mechanisms of meaning extension are usually termed metaphor and metonymy.For many cognitive linguists, metaphor is an all subsuming term for all cases of meaning extension.However, in this paper the two distinct terms will be preserved adopting the view that they are two poles in a continuum rather than separate categories (cf.Barcelona 2000, Radden 2000).The idea goes back to Jakobson (1956 mentioned in Barcelona 2000 andDirven 1993).
Metaphor is understood in terms of one domain mapped onto another (A mapped onto B), e.g. the expression to arrive at certain criteria is a metaphorical expression of the metaphor reaSoNINg IS a jourNey.In this paper, metaphors will be written in small capitals and will follow the pattern A (target) is/as B (source).
Metonymy, on the other hand, seems to work on the contiguity principle, when mapping/s occur within one domain, e.g. the expression the data suggest refer to the researcher suggesting that the data could be interpreted in one or another way (also see Dirven 1993).

Overall results
The present pilot investigation has identified 74 metonymical expressions including 27 expressions in LT and 47 in EN.Metaphors seem to be much more pervasive and amount to the total number of 317 metaphorical expressions occurring slightly more frequently in EN than LT (182 and 135 expressions respectively).
Metaphors seem to cover a broad range of source and target domains with the metaphors of laNguage (aNalySIS)/reaSoNINg aS a buIlDINg/Structure prevailing in both languages.Other major types in both languages would include laNguage (reSearch/learNINg) aS a meaSurable eNtIty, laNguage (learNINg) aS aN ecoNomIc actIvIty, laNguage (aNalySIS) / wrItINg aS a PerFormaNce/Play, laNguage (learNINg/reSearch)/ reaSoNINg aS a jourNey/ SPace/movemeNt.

Metonymical extensions
What in this paper is treated as metonymical expressions researchers of academic discourse have often discussed in terms of the combinability pattern inanimate subject+active verb.For example, Master (2001) has undertaken to investigate active verbs combined with inanimate subjects in different sciences: chemistry, biology, psychology, computer science and geology.Interestingly, his findings point out fairly high frequency of the so called verbs in the explanatory environment, e.g.show, indicate, suggest.In the framework of the present investigation most of these cases are metonymical expressions, e. g: (2) The paper will explore the ways in which writers use them [reporting clauses].
(3) A number of studies have proposed categories for classifying reporting verbs.
So, in the above examples (2) and (3) the metonymical expressions are based on the contiguity principle, which places the extension of meaning within one and the same domain, e. g. paper/study/ research and its author etc.
Interestingly, in terms of frequency of active verbs used with inanimate subjects the present investigation seems to have identified similar tendencies to Master's.So the verbs show and suggest in the EN data (out of 47 expressions 7 and 5 cases, respectively) have been identified among the top three most frequent verbs.The LT corpus, which seems to resort to metonymies to a much smaller extent, still has the verb rodyti (show) as the most frequent (8 cases in 27 expressions).LT also favours verba dicendi: sakyti (say, tell), aptarti (discuss) and teigti (assert, claim).
(5) The paper suggests a number of possible future lines of enquiry.
(7) Daugelis [studi]jų nedaug ką tepasako ir apie užsienio kalbos prigimtį... <Most studies tell hardly anything about the nature of a foreign language...> Interestingly, the verb suggest so frequently encountered in EN academic texts has hardly any equivalents in LT.This might be concerned with its function as a hedge in AD and culture-specific strategies of hedging in EN and LT.Hedging devices in LT academic discourse seem to favour various forms of galėti (may, can) and has nothing similar to the EN suggest (see Šeškauskienė 2005).The LT data has also manifested an interesting feature which might have important implications for further study in the field: some authors have not used a single metonymical expression at all.This tendency has also been observed in my oral discussion of the issue with some students of Lithuanian in Vilnius University.They tended to disagree on the acceptability of utterances of the type: the paper discusses.This seems to confirm my intuition that the LT academic language is presently undergoing changes, possibly under the influence of English, and there are quite a few innovations that have not been codified as yet but soon might find their way into the language system.The intuition, however, has to be verified on a larger amount of data.

Metaphorical extensions
l a N g u a g e (a N a ly S I S )/r e a S o N I N g a S a b u I l D I N g / S t r u c t u r e The above metaphor seems to be equally instrumental in both languages.The analysis or the language is seen by investigators as a multi-layer structure, they construct claims, knowledge or even objectivity and stance.Lithuanian seems to favour the foundation or the buttresses in the overall structure of language or research, which is why the majority of linguistic expressions make use of the words like atrama, remtis, pamatinė/bazinė struktūra, pamatas.English, on the other hand, gives a fuller picture of the "construction site" and covers such stages as design, construction and consolidation, hence the key elements of linguistic expressions include construct (objectivity, impersonality, claims, knowledge), base/basis of (arguments), (knowledge) building, consolidate (the argument) and the like.Neither LT nor EN seem to favour other elements of a building, such as the roof, windows or balconies.More often reference is made to rather abstract layers or levels, constituent elements or supporting arguments.Interestingly, what often is constructed in English can be hardly thought of in terms of construction in LT.So tests are designed, knowledge is built or constructed, objectivity is constructed only in EN; in LT tests are made up, knowledge or objectivity created.l a N g u a g e (r e S e a r c h / l e a r N I N g ) a S a m e a S u r a b l e e N t I t y Interestingly, both languages seem to be equally prone to conceptualising linguistic investigation or language learning, or any other language-related activity in terms of a measurable item or entity.The measuring axis extends vertically and horizontally; language /research/ learning can be measured in terms of quantity, frequency, weight, on a scale and within certain limits.Thus capabilities are measured in levels, the language use either in levels or in width (extensive use), some items manifest high or low frequencies, the vocabulary is measured as extensive; however, knowledge is mostly extensive in EN but deep in LT.Also LT describes research as weighty (e. g. svarūs tyrimai), which is hardly possible in EN.Depending on the topic of research, language/research is thought of in terms of more general (e. g. at the end of the scale, similar level of use) or more specific (e.g. high occurrence of ARGUE verbs, seven principal dimensions of knowledge) measurements.l a N g u a g e (l e a r N I N g ) a S a N e c o N o m I c a c t I v I t y Economic thinking seems to have long ago crept into our life and language.Therefore, rich/poor morphology or vocabulary, language acquisition, knowledge output, a high demand of translators, possessing knowledge, productive models, patterns or theory and language production errors do not strike us as unusual and have helped model this metaphor in both languages.The metaphor has generated very similar expression in both languages; presumably, quite a few of the above patterns have come into Lithuanian from English.Therefore, more or less the same conceptualisation has been preserved.l a N g u a g e (a N a ly S I S ) /w r I t I N g a S a P e r F o r m a N c e / P l ay The metaphor is a natural extension of our thinking about wrItINg aS a Stage where the author is seen as original, producing his/her own ideas and using language as an instrument of his ideas.Therefore, the 'stage' metaphor seems to have extended into other languagerelated activities, like linguistic research or writing about research, or just language.Most EN or LT speakers are familiar with the metaphorical role of words, sentences or any other elements in any context, or the role of the writer or researcher.Though EN texts seem to be more prone to the above metaphor, combinability patterns with the role are almost infinite in both languages.Interestingly, in EN the role can sometimes be replaced by visibility or presence, or even more terminological stance (e.g. the writer's stance).So the writer can be more or less visible in the text; s/he can exercise lower/higher visibility.The latter case manifests one of multiple cases of blending when two or more metaphors or some elements of source and target domains merge.So a higher visibility of the author merges the wrItINg aS a PerFormaNce and laNguage aS a meaSurable eNtIty metaphors.The construction of the writer's stance also involves the buIlDINg metaphor and the Per-FormaNce metaphor.The above instances only partially demonstrate the complexity of the problem, because in language clear prototypical cases are not always more frequent than peripheral or merged.The above metaphors are not very numerously represented; however, they are close in nature and sometimes overlap.It is quite logical that whatever moves in space can be treated in terms of a journey; however, what moves or is located in space does not necessarily is performing a journey.So in LT, you can move closer to the question or problem meaning that you are likely to answer or solve it, like in the following: (8) einama prie klausimo, grįžtama prie klausimo, prieita prie išvadų <go PassPraes to a question, return PassPraes to the question, go PassPast to conclusions> In EN, you can arrive at certain criteria, we can be a long way from being able to (…) predict tendencies.Both arriving and being a long way from signal the jourNey metaphor.Spatial (proximity) metaphors usually exploit collocations with field which is why we have semantic fields, field of academic discourse or researchers working in the field.LT, however, tends to use area in the general meaning and field is restricted to terms like semantic fields (semantiniai laukai).Another fairly recent pattern of thinking in LT and EN is in terms of centre and periphery, like in funkcijų centras ir periferija, peripheral use/ functions etc.
The movement metaphor is not easily identifiable since it often merges with the journey metaphor.However, in most general cases concerned with raising a question or an emerging problem/feature, or the author moving on to consider other issues, the metaphor of movement is most obvious in both languages.

Minor types of metaphors
Though scarcely represented, they include a variety of source and target domains.Thus language or learning can be thought of in terms of a coNtaINer where something is put, e.g.: (9) (…) writers can incorporate their attitudes and judgements into the text.
(10) Mokymasis -tai žinių perorganizavimas, o ne tik jų laikymas atmintyje <Learning is reorganising the knowledge rather than keeping it in one's memory> In EN, the laNguage aS a coNtaINer metaphor is often signalled by the word incorporate; in LT, the word talpus (capacious) combined with definition or text seems to be indicative.
Other rather sporadic metaphors include laNguage/reSearch aS a humaN beINg, laNguage aS aN emotIoNal value, laNguage aS a Flow, laNguage aS a DISguISe/cover For a Secret, orDer aS a vehIcle IN ItS trackS, reSearch aS lIght, lINguIStIc aNalySIS aS a PIcture, lINguIStIc aNalySIS aS war/comPetItIoN etc.Interestingly, the metaphor laNguage aS aN emotIoNal value is exclusively confined to Lithuanian, which is often seen by language purists as something despised and neglected, not loved etc.The scarcity of the metaphor lINguIStIc aNalySIS aS war seems to confirm the initial assumption that the language of research is not as aggressive and adversative as political discourse.

Conclusion
The present paper has attempted to verify the hypothesis that academic discourse follows the tendencies of meaning extension present in any natural language, despite an overall attempt/inclination of the language of research at rationality, disambiguation and monosemy.This is partly due to the nature of the selected discourse-in linguistic texts language serves as an object and an instrument of study.
The EN and LT data collected for the present investigation have shown marked tendencies of the texts to express ideas through two major mechanisms: metaphor and metonymy, the latter being more than four times more productive than the former.Metonymy has interestingly been given preference by some authors and absolutely evaded by others.
The metaphor has manifested a variety of types.The most frequent have focused on language or any language-related activities thought of in terms of building or a decomposable structure, a measurable entity, an economic activity, as a performance/play, a journey or space, or movement.Less frequent types have included language thought of in terms of a human being, an emotional value, a flow, a disguise for a secret, order as vehicle in its tracks, research as light, linguistic analysis as a picture or war etc.
All major patterns have been represented in both languages manifesting some language/ culture-specific peculiarities of realization and combinability.Some minor patterns (like laNguage aS aN emotIoNal value) have turned out to be rather culture specific.
l a N g u a g e (l e a r N I N g / r e S e a r c h )/r e a S o N I N g a S a j o u r N e y , S Pa c e o r m o v e m e N t