This paper investigated the use of of, the most frequently occurring preposition, in essays written by 416 matriculants at a Korean university. The learner corpus consisting of these essays contained 1,250 tokens of of, and these tokens were first analyzed according to their functions, i.e., integrative, separative, idiomatic, and others, adapted from Lindstromberg’s (2010) categorization of the functions of of. To ascertain what types of errors were made regarding of, the 46 tokens of of identified as errors were further categorized as addition, misordering, misformation, or wrong position. Of these four categories, addition was found to be the most frequently occurring type of error (31 of the 46 tokens), while wrong preposition and misformation accounted for ten and three tokens, respectively. No tokens of misordering were found, however. Also noteworthy is the finding that, albeit occurring mostly in simple constructions and basic functions, of was used accurately in over 95% of all the tokens found in the learner corpus.

When one is reading or writing, various forms of bottom-up knowledge—e.g., grammar, vocabulary, orthography, and the mechanics of written language—interact with the top-down processes at the level of discourse. One of the reasons English language learners encounter a lot of difficulty when they are engaged in academic reading and writing is that they have been taught grammar only at the sentence level and not at the discourse level. In order to illustrate how discourse-based grammar instruction can facilitate the acquisition of academic reading and writing, this paper discusses various ways in which the following four topics can be dealt with at the discourse level: cohesion, the tense and aspect system, comma usage, and existential there. Obviously, discourse-based grammar instruction should be integrated with the teaching of the other top-down and bottom-up skills necessary for academic reading and writing. EFL teachers, however, need to know more about discourse grammar to effectively make their learners aware of it and to offer them learning activities that will contribute to better reading comprehension and written production.

This study investigates the types and frequencies of verb errors in Korean college students' essays in order to ascertain what aspects of English verb use Korean learners find most troublesome. The data used in this study were retrieved from a learner corpus consisting of essays written by 399 students who major in humanities at a university in Seoul. The 686 verb errors found in the corpus were classified into the following four major categories: (a) omission of necessary items in a verb phrase, (b) addition of unnecessary items in a verb phrase, (c) misformation ofa verb phrase, and (d) misordering of items in a verb phrase. A careful examination of these 686 verb errors has revealed that misformation is the most common form of error, accounting for over 60% of all the errors. A sub-category of misformation errors, agreement errors in turn accounted for more than half of all the 4 I 6 misformation errors (216 tokens), a number bigger than any of the other three categories of error types, i.e. omission (175 tokens), addition (72 tokens), and misordering (23 tokens). This finding might have resulted from negative influence from the students' L I, as Korean verbs do not conjugate according to grammatical person. Another noteworthy finding is the fact that the students made a great number of errors with both the lexical and the auxiliary uses of be and have. Considering that both uses of these two verbs are taught early on in Korea, this finding suggests that Korean students need to be continually provided with contexts in which they can practice different uses of be and have.

This study reports on the findings of a corpus-based analysis of Korean college students' use of English conjunctive adjuncts. Unlike many previous studies that mainly focused on describing the position and semantic types of conjunctions used by learners of English, this study examines grammatical errors of conjunctive adjuncts found in a leaner corpus that consists of 102,632 words written by 399 Korean college freshman students. The main findings of the study can be summarized as follows. First, learners tend to use sentence-initial coordinators even when the sentences before and after the coordinators are not long enough to warrant such usage. Second, sentence fragments occur much more frequently than run-on sentences with the 10 most frequent conjunctive adjuncts found in the corpus. Finally, learners often add unnecessary punctuation marks or omit necessary ones after conjunctive adjuncts, errors which sometimes make it difficult for readers to understand the text. All these errors amount to the conclusion that many Korean learners of English at the university level lack the necessary grammatical knowledge of English conjunctive adjuncts to use them correctlyin academic writing.