A Brief Literature Review of Studies on Lexical Chunks at Home and Abroad

This paper attempts to offer a general review of studies on lexical chunks at home and abroad to reveal prospective readers a clear path to approach this promising language phenomenon. It starts with definitions of chunks, followed by an exemplification of the essential characteristics, different classifications, and functions proposed by researchers, and then it summarizes some major findings both theoretically and empirically. Based on these, some crucial aspects are pointed out that previous studies fail to cover but a worthwhile attempt for the future study.


Introduction
In the 21th century cognitive psychology has developed so much and achieved fruitful findings which offer a new angle for the field of cognitive linguistics and those linguists come to realize the fact that language essentially consists of a large number of ready-made chunks. People in everyday interaction will consciously or unconsciously turn to those well-prepared multi-word units for immediate encoding, with no reference to using some complicated grammatical rules or principles. Lexical chunks can be used to realize the interchange of semantics, syntax and pragmatics, spur the progress of syntactic rules, simultaneously assist language learners to achieve authenticity and appropriateness of their expressions, and therein promote the language output effectively. On the other hand, lexical chunks serve as a medium between structural approach and communicative approach, which can not only highlight grammar, but also lay great emphasis on the proper usage of the local context. Virtually, the essence of the lexical approach lies in this wise statement by Levis"language consists of grammaticalised lexis, not lexicalised grammar" (1993, p. 89). It can be interpreted in this way: Learning a language is far from mastering these grammatical rules or memorizing thousands of independent words but often multi-word ready-made chunks, which is beneficial to language learners. This paper attempts to offer a general review of the development of studies on lexical chunks at home and abroad to reveal prospective readers a clear path to approach this promising language phenomenon. It starts with definitions of chunks, followed by an exemplification of some essential characteristics, different classifications, and functions proposed by researchers, and then it summarizes some major findings both theoretically and empirically. Based on these, some crucial aspects are pointed out that previous studies fail to cover but a worthwhile attempt for the future study.

Theoretical and Practical Significance
For decades, countless researchers, linguists, and even language teachers have been exploring the nature of language based on their respective specialized field, from Saussure to Chomsky, Austine to Grice. It is well accepted whichever language consists of a series of rule-controlled elements, which linguists have been sparing no efforts to uncover. There is no doubt that a decent knowledge of grammar plays an essential role in guaranteeing a continuous development of target language learning to some extent. Nevertheless, grammar-learning does not necessarily represent the language-learning as a whole because a number of language learners can still not communicate with others in speaking and writing although they have already remembered a lot of grammatical rules. Indeed, grammar serves lexis (Widdowson, 1993). The latter is central to language. Hopefully, by means of exploring the role of chunks played in children' language development and adult language use respectively, linguists have come to realize the dual features of language. They, on the one hand, still acknowledge the human superb capacity to utilize rule-based analytic knowledge of language in building up new sentences; on the other hand, they also highlight learners' sharp memory for semi-structured prefabricated chunks. Bolinger (1976) notes that language users tend to check those existing ready-made chunks without constructing utterances each time, instead of construction with a rule system, Furthermore, one of the commonest characteristics of SLA, or even the native speakers, is that language learners intend to employ masses of unanalyzed prefabricated chunks in predictable social contexts. It is suggested that a good command of unanalyzed prefabricated chunks are essential to language acquisition, and learning. Practically, many researchers through examining language learners' output, recognize some features of those ready-made chunks that they used, and some recent studies have concentrated on carrying out corpus-based investigations into the relation between collocational usage and writing ability in native and non-native speakers of English (Cowie & Howarth, 1996). For example, Cowie (1996), who was particularly interested in inter-lingual use of collocation, and carried out a comparative analysis of written performance. One of her significant findings is that foreign language learners tend to use fewer prefabricated patterns than the native-speaker counterparts do. That maybe will explain the deficiency of those foreign language learners.

Research Methodology
This paper will apply a kind of qualitative method to approach this promising language phenomenon. In particular, the paper will first investigate the current academic study of lexical chunk lexical chunks at home and abroad in terms of definitions of lexical chunks, essential characteristics of lexical chunks, different classifications of lexical chunks, and then summarize some major findings both theoretically and empirically. Finally, some crucial aspects are pointed out that previous studies fail to cover but a worthwhile attempt for the future study.

Definitions of Chunks
Becker and his colleague coined and defined the term chunk as early as in the 1970's, but it has come into a broad sense and widely accepted and interpreted by different experts from their respective academic field. As a matter of fact, a diversity of terms are applied to refer to much the same text phenomenon, ranging from fixed terms like idioms and proverbs to loose connections like collocations. These terms are often referred to, such as ready-made language, formulas, fixed or semi-fixed expressions, lexicalized sentence stems, lexical phrases, prefabricated chunks.
It is easy to see that this language phenomenon appears to be so confusing that Jespersen (1924) is the first linguist to make a general distinction between "formulas", and "free expressions". He points out that free expressions are made in each case, but formulas are extracted from one's memory. Furthermore, he claims that "a language would be a difficult thing to handle if its speakers had the burden imposed on them of remembering every little item separately." (1924, p. 76). Brown (1980) just follows his predecessor and coins the term prefabricated routines, to describe utterances that are memorized as a whole. It is Wray who employs the term formulaic sequence, and "sees it as a continuous or discontinuous sequence of words or other meaningful elements, which is or appears to be prefabricated; that is stored and retrieved whole from memory at the time of use, rather than being subject to generation or analysis by the language grammar." (Wray, 2000, p. 465) Virtually, each researcher tends to select an appropriate term for this linguistic phenomenon to fit in his or her specific research perspective. No matter what kind of terms they choose, however, it can be safely said that the differences between them are more a matter of degree than of kind. For the sake of clarity in this research, the term of lexical chunks is chosen. The paper refers to lexical chunks as special multi-word items, between grammar and vocabulary, and fixed or semi-fixed, which consist of a sequence of two or more words, semantically and/or grammatically being a meaningful and inseparable unit.

Different Categorizations of Chunks
It is Becker who attempted to categorize lexical chunks into sub-types in 1975. He renders the categorizations of lexical chunks via adult native speakers into five sub-types as is shown in Table 1 One demerit of Becker's classification is that he fails to distinguish between form and function. What's more, he ignores a fact that chunks are featured with variability. A chunk is more likely to be fixed only with one part but still consists of an open slot in which either a word or phrase is filled if needed. Despite this, his classification lays a good foundation for the the later researchers.
Nattinger and DeCarrico's taxonomy is well-known in the Chinese academic community. They use lexical phrases to refer to chunks and see them as "conventionalized form/ function composites that occur more frequently and have more idiomatically determined meaning than language that is put together each time." (Nattinger et al., 1992, p. 25) They divide lexical phrases into four categories according to form, as is displayed in Table 2. More importantly, they also cross-categorize lexical phrases according to the function, as is shown in Table 3. From all the categorizations mentioned above, it is not difficult to come to the conclusion that there is still no fixed standard for chunk classification and researchers set up their own systems for their own research purposes. This paper will prefer to adopt Nattinger and DeCarrico's taxonomy as the theoretical basis for the following analysis.

Functions of Chunks
As we all know that lexical chunks play an essential role in acquiring and learning a language. But it still appears a confusing issue how it works in practice. Hopefully, it is Moon (1998) who is the pioneer to theorize its five functions according to the way they contribute to the content and structure of a text. Informational function: it suggests that lexical chunks state proposition and convey information. E.g.by means of, in the running. Evaluative function: it shows the lexical chunks may convey speaker's evaluation and attitude. E.g., it is an ill wind, a pain in the neck. Situational function: the lexical chunks are related to extra-linguistic context and respond to situation, e.g.excuse me! Modalizing function: it means that lexical chunks may convey with values, advice, requests, etc. e.g., you know what I mean, etc. Organizational function: It says that lexical chunks have a function of organizing text and signaling discourse structure, eg, by the way, for example. Based on the functions mentioned in literature, this thesis would rather elaborate its functions from the following aspects.

Alleviating the Burden of Language Encoding
Chunks may save learners' time and energy, and facilitate learners' concentration in the process of language encoding because chunks are extracted and processed as a whole, which speeds up language processing significantly. Becker mentions chunks"give us ready-made framework on which to hang the expression of our ideas, so that we don't have to go through the labor of generating an utterance all the way out from we want to say anything." (1975, p. 17). A phenomenon can be frequently witnessed that speakers would express fluently when talking about familiar experiences or activities in familiar phrases, which suggests that people are dependent heavily of ready-made chunks to save their energy and time in language processing. Since chunks are stored as a whole, they require no more decoding effort than big words (Ellis, 1996, p. 111). This can be best illustrated by college English learners experiences in China. When a sentence, like "Do you want cash back please?" is heard, one needs much more time than native speakers to react. This kind of sentences or chunks is usually split into pieces and encoding again for understanding the whole meaning. Unlikely, native speakers quickly process this information as a whole. Chunks, undoubtedly, save their time in processing.

Facilitating the Social Interaction
It is universally acknowledged that chunks may be not the only ways for expressing an idea, but they are undoubtedly very good ways for expressing. Krashen (1978, p. 2) suggests when learners are compelled to speak before they are ready, they tend to apply those formulaic patterns to make up for the deficiency of their language knowledge. In other words, if the learners have no adequate grammatical rules and vocabulary to make appropriate speeches, they would turn to read-made chunks in order to meet the basic communication needs. Therefore, when language learners or users have the least language knowledge, using chunks is just like an effective tool to maintain the social interaction. For example, people may need to use chunks for daily requesting, shopping, praying or other social-related frames to adapt to the society.

Evidence from Corpus Linguistics
There is no doubt that information technology has shaped the world we are living in, in particular, the medium by which we see the world: language, and, importantly, it has definitely offered us new approaches of accessing language, and has already revealed "quite unsuspected patterns," which "traditional descriptive frameworks are normally not able to account for" (Sinclair, 1991, p. 29), and via which a large number of naturally utterances would be analyzed.
As early as the year of 1961, an corpus of oral utterances were approached by Jones and Sinclair as the first bold trial to track the evidence of lexical instead of grammatical structure to show the link between words which were elements of the frequently recurring structure of the language and their related meaning. In spite of the fact that Jones and Sinclair can not approach adequate sum of data resources with plausible methods, they were capable of tracking convincing proof: the sequence of words concerning areas of meaning; the delexicalisation of some high-frequency words; and certain patterns of collocation. Since then there have been growing countless probes into the collocation tendencies of individual words via computer corpora.

Chunks in Second Language Acquisition
Nattinger & DeCarrico (1992) point out that language is first picked up as undivided chunks, and then split into its organic segments, which could become accessible and valuable for later use in principle-controlled language interaction when it comes to both the first language and the second language acquisition. Echoing the above-mentioned assumption, Skehan (1992) reckons that language tends to be memory-based and idiomatic. He furthers his argument that language learning is made up of three different periods: lexicalization, syntacticalization, and relexicalization. The proceeding progress of the first language acquisition starts with the process of lexicalization, then sequences of lexical segments become syntactic, called syntacticalization, and finally, the more the language system thus developed, the more it would become "relexicalized''. By the means of "relexicalization" Skehan holds that the language learners have to learn the way to cope with the flexible and energy consuming syntactic pattern in the case of information processing requirement. This paper contributes to an essential point that the core of grammar in linguistics has been already practiced by many SLA studies, but quite few studies are paid adequate attention to inter-language vocabulary. Those existing vocabulary acquisition studies have only attempted to concentrate on the role of memorization and repetition of every single and separated word in the second language, and various learning strategies and skills are also short to supply. The over-emphasis on grammatical analysis and rule-formation in language acquisition also, certainly, veils the fact that it is not beneficial to analyzing those undivided words or expressions.

Chunks in Second Language Teaching Theories
Those Linguists and researchers, who are the typical representative of putting the Lexical Approach in class, such as Lewis, Nattinger, & DeCarrlco, gain considerable benefits or even enlightenment from some sensational findings of corpus linguistics. Lewis (1993), chief representative of this approach, speaks highly of the meaning-fixed, well-prepared multiword chunks. He furthers his idea that language is not only composed of fossilized grammatical points and vocabulary but frequently of a considerable sum of multi-word lexical chunks. In addition, teachers who apply the Lexical Approach should not waste too much time in dealing with the language structures in the classroom, but are supposed to guide students' attention to these practical and well-prepared chunks: in the practice of language teaching, the importance of noticing as well as listening and the value of those useful chunks should be highlighted. The basic aim of the lexical teaching activities should be interactive, awareness-raising of lexical chunks, rather than a prim and deficient teaching in a high-anxiety and prim learning environment. However, till today, there are quite few works or findings on how to put the lexical approach in the authentic classroom. Most of them just focus on some theoretical aspect, the more detailed and practical attempt is urgently in need.

Research Methodologies
It is just assumed that the use of chunks plays an essential part in the process of language learning, but scientific evidence is still in need to justify the assumption, for which many studies have been undertaken. These studies tend to be carried out under the categories of quantitative and qualitative one.

Quantitative Studies
A majority of early studies, both quantitative and qualitative, pay attention to collocations, chunks fixed with one part and to fill up, but data or statistics are commonly involved in quantitative studies, and the qualitative studies do not necessarily employ statistical procedures. Actually, these studies are reckoned as quantitative mainly because they all prefer to apply figures to illustrate the deficiency or shortage in the knowledge of collocations to backup the importance of teaching and learning collocation. For example, Channell (1981) undertook an experiment where eight ESL students with C2 level are required to complete a collocation quiz: matching 4 adjectives with 16 nouns. The result shows an embarrassing situation. Although the eight students know very well each single word, they fail to match 111 acceptable collocations, and 24 unacceptable collocations are made. Based on the results, Channell points out that teaching a new word is teaching how the word collocates with other words in exact contexts. Zhang (2004) undertakes a survey on 15 college students who are divided into three different linguistic groups. This survey attempts to display the relationship between the use of chunks and students' oral English fluency. The results suggest that more advanced the English learners' English proficiency is, the more chunks would be employed by them, and their oral English would be more communicative and authentic. One thing needs to be recognized that the above-mentioned quantitative studies apply the limited samples or participants, which would impair the validity of the study. Still, quite few researchers pay attention to kinds of formulaic expressions aside from collocations.

Qualitative Studies
Yumoto (1992) spends two years in observing the usage of lexical chunks of her two Japanese-speaking kids, aged four and eight. She witnessed the process of undivided utterances being segmented and pieced into new patterns and considered the process as"unconscious pattern practices and substitution exercised" (1992, p. 1). She concluded that the utilization of formulaic expressions lessens the pressure of learning but boosting one's communicative ability and the natural process of acquiring the formulas assists the learner to form the internalized rules of the language. The strong point of the above study is that children at a very young age are used as their subjects to suggest an underlying fact that children acquire chunks and use them long before they know anything about their structure and before they can create any novel sentences of their own. A majority of qualitative studies choose small children as their participants who are likely to acquire the second or even the third language easier. Thus the practicability of their findings would be doubted because the process of young children' language acquisition would be different from that of the adult, to some extent. Still, many qualitative studies have just draw on the process of oral skills, while ignore the rest of other language skills, such as reading, listening, and writing.

Conclusion
To summarize, the literature review above in terms of lexical chunks can reveal prospective readers a relatively clear path to approach this promising language phenomenon. Compared with the researchers abroad, Chinese researchers appear to pay attention to this field later than his counterpart, and a large proportion of academic findings with regard to lexical chunks in China tend to be basically theoretical introduction to the role of chunks in language learning. Some key points can be also draw up. Firstly, definitions of this term are still varying, and different researchers tend to approach this term for the sake of their personal research perspective. Secondly, there still exist different classifications for this term based on different standards, and some of the classifications appear somewhat overlapping. Thirdly, the role played in the first or second language acquisition is obviously www.ccsenet.org/ells English Language and Literature Studies Vol. 6, No. 2;2016 active, but it is still in doubt whether it plays the same role in EFL, like English learning in China. Fourthly, the research methods tend to be oversimplified, either qualitative or quantitative, and a better researching method is still in need. Last but not least, even though it is well proven that lexical approach can be beneficial to the language teaching, but till today quite few studies or works specify the practical procedures of the lexical approach in class, let alone the listening, speaking class, and writing class. Therefore, it is not persuasive to say that the results can be well helpful for language learning.