RESUlTIVE CONNECTORS IN aDVaNCED lIThUaNIaN lEaRNERS ’ ENglISh wRITINg

The importance of the analysis of learner language has long been recognised. The ICLE project has facilitated the research into interlanguage offering reliable data on advanced students’ written work. The Lithuanian subcorpus of ICLE is currently being compiled at Vilnius University. The present paper will report on a pilot study carried out on the material of LICLE on the use of resultive connectors and also will present a brief comparison of the Lithuanian learners’ usage with that of native speakers’ as well as advanced learners’ of other language backgrounds. The central research questions will be the following: • whether advanced Lithuanian learners use English resultive connectors in the same way as native English university students; • whether advanced Lithuanian learners’ usage of English resultive connectors is similar to other non-native learners’; • what possible explanations could be offered.


Introduction
The importance of the analysis of learner language has long been recognised.The ICLE project has facilitated the research into interlanguage offering reliable data on advanced students' written work.the Lithuanian subcorpus of ICLE is currently being compiled at Vilnius University.
The present paper will report on a pilot study carried out on the material of LICLE on the use of resultive connectors and also will present a brief comparison of the Lithuanian learners' usage with that of native speakers' as well as advanced learners' of other language backgrounds.The central research questions will be the following: whether advanced Lithuanian learners use English resultive connectors in the same • way as native English university students; whether advanced Lithuanian learners' usage of English resultive connectors is simi-• lar to other non-native learners'; what possible explanations could be offered.• Numerous studies have been carried out on the topic of connectors in learners' written discourse.Granger and Tyson (1996) looked into the use of connectors in essays written by the French, Altenberg and tapper (1998) -Swedish, tankó (2004) -Hungarian, Narita, Sato and Sugiura (2004) -Japanese, Fei (2006), tang and Ng (1995), yaochen (2006), Milton (2001) -Chinese, yoon (2006) -Korean, Chen (2006) -taiwanese, Leńko-Szymańska (2007) -Polish and other students, etc. While Chen (2006), tankó (2004), Yoon (2006) and Milton (2001) reported on a general overuse of connectors by non-native speakers of English, altenberg and Tapper (1998) found a general tendency for underuse.Granger and Tyson (1996), on the other hand, noticed neither general overuse nor underuse of connectors in students' essays.Despite the observed differences in the general use, overor underuse of individual connectors was observed by all the authors.
Connectors have been analysed by their semantic categories (altenberg 2007), the place they occupy in a sentence (Smits 2002(Smits , altenberg 2006)), etc.Though there are but a few studies carried on following the latter criteria, attention to the sentence position is paid in the majority of works on connectors.Cross-cultural problems of connector usage have also been addressed.(Ventola 1992).the use of connectors in the Lithuanian students' written work is a new area and has hardly been addressed yet (Burneikaitė 2007).

Material
For the purpose of this study a sample corpus of LICLE totalling 114,855 words was used.For comparison of resultive connectors, the native speakers' data was taken from the British segment of Louvain Corpus of Native Speakers of English.It consists of 95,695 words.the findings from LICLE are also compared to the findings from other learner corpora reported in numerous studies.References to the studies are given when they are citied.

Methodology
CIa (Contrastive Interlanguage analysis), as a type of Ca (Contrastive analysis), is used in this paper because it "establishes comparisons (...) between native and learner varieties of one and the same language" (Granger 1996: 43).Since the material is taken from a comparatively small corpus, according to Sinclair (in Flowerdrew 2004: 17), CA is the best research method as comparison allows us to notice differences regardless of the corpora size.

Findings
Following Quirk's et al (1991) classification, connectors can be grouped into seven semantic categories.The data from LICLE and LoCNESS indicate that the four most commonly used semantic categories are the same in both corpora.The semantic category under consideration occupies the third and the second position respectively.However, despite the lower position observed in LICLE, there is only a slight tendency for the Lithuanian learners to underuse resultive connectors.Moreover, this difference is not statistically significant (LL -1.02).It is worth noting that a similar tendency of underuse was observed in SWICLE (Swedish subcorpus of ICLE) (altenberg and Tapper 1998: 85) and in the Hungarian corpus, which was compiled following ICLE criteria (tankó 2004:171).one of the possible explanations for this underuse is offered by several authors (altenberg andTapper 1998, Basturkmen 2002), who claim that resultive relations can rather easily be inferred from the context and, therefore, the omission of connectors belonging to this semantic category (Example 1) is less risky than in the case of other connectors (Example 2).
(1) Since the superiority of English is manifested by its use in various areas of life, it may decrease the people's self-esteem and trust in their native language.Therefore, it may be tempting for them to abandon their language in favor of the seemingly more valuable English.(ICLE-LT-VI-0028.1) (2) It is like a journey or an adventure that reveals who you are.On the contrary, writing per se could be perceived not only as a positive process, for it engages a lot of hard work.(ICLE-Lt-VI-0046.1) The phenomenon of underuse, however, should not be considered to be uniform for the interlanguage.Chinese students (yaochen 2006: 34, tang andNg 1995: 108), contrary to the Lithuanian, Swedish and Hungarian learners, tend to use more resultive connectors than the native speakers.
A closer look at individual resultive connectors conforms to Granger and tyson's (1996) findings that though there is no significant overall overuse or underuse of connectors in learners' writing, there is a strong evidence of over-or underuse of individual connectors.In LICLE, only two resultive connectors, namely, therefore (LL -16.45) and thus (LL 9.53), are used with a significant difference.Despite the fact, that therefore is significantly underused, it, together with the overused thus, is among the ten most frequent connectors in the Lithuanian students' essays (table 1).The data from Table 1 show that therefore and thus appear in all the corpora under consideration and occupy rather high positions, the span being between two and six.It could be explained by the fact that the greatest majority of texts in the corpora are argumentative essays, which, according to tankó (2004:167), can be characterised as having syllogistic sequence premise 1 + premise 2 therefore conclusion: (3) Such a society would not borrow words, ideas, and culture from other nations.It would be not change its way of living and, therefore, it would not alter its language and mentality.... (ICLE-LT-VI-0087.2) It is interesting to note that LICLE differs not only from LoCNESS but from all the other corpora mentioned in Table 3 as well.Though in all the corpora therefore and thus are among the most frequent connectors, only in LICLE it is preceded by thus.Whether the Lithuanian learners should be encouraged to change the observed is a question for qualitative analysis.on the quantitative dimension, however, the Lithuanian students' results seem to be closer to the results from BYU-BNC: The British National Corpus academic section than to those of LoCNESS (Table 2).

Therefore
Therefore is the only resultive connector that is significantly underused by the Lithuanian learners.Its text and sentence position will be looked at in greater detail.
Table 3 shows the distribution of therefore throughout the text in LICLE and LoCNESS.It includes the number of all the occurrences in each position.

Ta b l e 3 . Distribution of therefore throughout the text in LICLE and LOCNESS
In both corpora, therefore is distributed almost evenly throughout the text with the exception of the middle-paragraph position, i.e. the body paragraphs of the text.though it is the most favoured position in both corpora, the Lithuanian learners significantly underuse it .this phenomenon could be explained by the general underuse of the connector under consideration.
Rundell and Granger ( 2007) noted that, in learner language, there are many cases when "an expression that carries the 'right' meaning is used in a non-preferred position in a sentence."this could clearly be observed in the use of therefore (table 4).
In the Lithuanian learners' essays, therefore is significantly overused in a non-preferred sentence-initial (SI) position (LL 21.98) whereas it is underused in a SM (sentence-medial) position (LL -67.22).The tendency to use connectors sentence-initially is not a feature characteristic only of the Lithuanian learners of English.It could be considered to be a common phenomenon of interlanguage (cf. tankó 2004: 176;Granger and tyson 1996: 24;Narita et al. 2004Narita et al. : 1174, etc.) , etc.) possible explanations for the observed differences adequate use of connectors is one of the ways to ensure cohesion.However, their misuse is an almost universal feature of interlanguage (Crewe 1990: 317).The reasons can be manifold.Language transfer patterns undoubtedly play an important role.
Another reason is insufficient and inadequate information presented in dictionaries and grammars (cf. Granger and tyson 1996: 23).Dictionaries vary in defining the register of the resultive connectors.Definitions, on the other hand, are very similar or almost identical.an even more alarming thing is that reference tools provide learners with contradictory facts.While in 'Improve your Writing Skills' section of Macmillan English Dictionary for Advanced Learners (2007: IW13), it is noted that, in academic prose, therefore is used less often than thus, the Longman Grammar of Spoken and Written English (Biber et al. 1999: 887) provides the contrary information.

Conclusions
The present pilot study of resultive connectors in LICLE and LoCNESS has revealed the following findings: Though overt signalling of resultive relations is more common to the native speakers, 1.
the difference is not statistically significant and could be considered to be negligible.
Since the semantic category under consideration is characteristic of argumentation, this tendency can imply that, at least on a formal level, the Lithuanian learners use the same structure of argumentative essays as the native speakers.2. Therefore and thus are among ten most frequent connectors both in the learners' and native speakers' corpora.their positions vary between the second and the sixth among top ten connectors.The native speakers tend to rely on a closer set of connectors than the learners of English, therefore, the fact, that the connectors under consideration appear in both corpora, could be seen as a step towards native-like use of this semantic category.The Lithuanian learners differ both from learners of other language backgrounds and native speakers on the ratio of therefore and thus, which is closer to that of the British National Corpus academic section than to LoCNESS. on a quantitative level, this could be seen as a great achievement of the Lithuanian learners and teachers.a qualitative analysis, however, is needed to give an objective evaluation of this phenomenon.theLithuanian learners' pattern to distribute therefore throughout different paragraphs of the text is similar to the one of the native speakers'.In LICLE, the general underuse of this connector is due to the significantly underused middle-paragraph position.this fact could be taken into consideration teaching students to write argumentative essays.The sentence initial position of therefore, though it is not favoured by the native speakers, is preferred by the Lithuanian learners as well as learners of other language backgrounds.It can be considered to be an interlanguage phenomenon.Possible explanations for the observed differences can be attributed to features of interlanguage, developmental errors, inadequate information in reference tools, language transfer, etc.The results of the study should be seen as tentative because more detailed research will be needed before firm conclusions could be drawn.

Table 1 .
Therefore and