This volume showcases some of the latest research on academic writing by leading and up-and-coming corpus linguists. The studies included in the volume are based on a wide range of corpora spanning first and second language academic writing at different levels of writing expertise, containing texts from a variety of academic disciplines (and sub-disciplines) and of different academic registers. Particularly novel aspects of the collection are the inclusion of research that combines rhetorical moves with multi-dimensional analysis, studies that cover both fixed and variable phraseological items (lexical bundles, phrase-frames, constructions), and work that is based on corpora of English as an academic lingua franca. Going beyond merely summarizing their findings, the authors also discuss... what their research means for academic writing practice and pedagogical settings. The volume will be of interest to researchers, students, and teachers who would like to expand their knowledge of how academic writing functions and what it looks like in a variety of contexts.;;Many corpus-based studies focus on the use of academic vocabulary in journal articles and textbooks while creating vocabulary lists and argue for the teaching of such lists, but few examine how students employ those words in their writing. The present study investigates English as a Foreign Language (EFL) learners’ use of academic vocabulary in their writing assignments as they attend a dual-degree program in STEM fields in Tbilisi, Georgia. Over a thousand student papers from their General Education courses are examined for their use of academic vocabulary, using Gardner and Davies’ (2014) Academic Vocabulary List. Results show variation in students’ academic vocabulary use and quantity while progressing in their studies and as they produce the different text-types in their coursework.;;This study aims to investigate constructions of the verb give used by Korean EFL learners, along with an in-depth analysis of ditransitive use and L1 influences. The use of give in a Korean EFL learner corpus and a reference native speaker corpus is analyzed in terms of different construction types and their frequencies. In-depth analysis of ditransitive use of give was conducted using three factors: pronominality, weight, and semantic classes. L1 influence was discussed and a Korean language corpus was used to provide evidence to strengthen the arguments. This study concludes with the discussion of limitations, and future directions for understanding Korean EFL learners’ developmental trajectory of construction acquisition within the usage-based framework.;;With English being the global language of research, academic English as a Lingua Franca (ELF) research has gained wide recognition. While early research showcased the dynamic nature of spoken academic ELF, written academic ELF, a more recent focus of research, remains to be under-studied. However, not only are there known differences between speaking and writing, writing is also the dominant mode of reporting new research in academia. Situated within usage-based Construction Grammar (CxG), the present study aims to identify constructions in written academic ELF in comparison to American academic writing. Our methodological approach, which combined a key word analysis with systematic explorations of fixed and variable phraseological items in a corpus of ELF academic writing, leads to the identification of constructions that are characteristic of written academic ELF.;;Writing from sources is a necessity for academic writing, but it is especially challenging for First-Year Composition (FYC) writers (Horning & Kraemer 2013; Serviss & Jamieson 2017). Yet, few studies have empirically investigated the source-based language use features of FYC writers. For this study, English as a first language (L1) FYC students ( N = 232) wrote in-class essays using two source readings. A multi-dimensional analysis on 13 linguistic features associated with the presence or absence of source use (e.g., reporting verbs, titles and authors, personal pronouns) was conducted, resulting in three dimensions, (1) Source-Based Concept Density vs. Prompt-Based Freewriting, (2) Impersonal Extension of Source-Based Concepts, and (3) Source Text Deixis. Implications of these dimensions for source-based student writing are discussed.;;This chapter presents the results of an investigation into the interaction between discipline and writing experience on lexical bundle (LB) production. The study compared the production of 4-word LBs by novice and expert writers in medical and non-medical fields. The findings indicated that a combination of both writing experience and field-specific knowledge led to an increased use of medical LBs. Field-specific knowledge and experience seemed to be a prerequisite for the production of medical LBs with interactions demonstrating that the use of lexical bundles increased with the combination of field-specific knowledge and experience. The chapter interprets these findings based on previous research of learners’ acquisition of LBs and discusses their implications for both research and pedagogy.;;We carry out comprehensive form – function mapping in Introduction-Methods-Results-Discussion/Conclusion – structured research articles across 30 academic disciplines by merging move analysis (revealing rhetorical structure) and multi-dimensional (MD) analysis (modeling patterns of linguistic variation). These two analytic paradigms converge to map the communicative functions of text segments with patterns of functional linguistic variation, but the juxtaposition of these two approaches requires adaptations to the traditional MD methodology. The primary purpose of this chapter is to provide a detailed account of the methodological complexities of combining MD analysis and move analysis. We briefly present the results of the MD analysis, acknowledging that further methodological adjustments may be needed to arrive at an optimal multi-dimensional description of moves.;;This chapter reports on a study that examined over 700 four-word lexical bundles (Biber et al., 1999) identified from a multi-disciplinary corpus of over 25 million words from introductory courses across nine disciplines. Two types of bundles were identified: cross-disciplinary (occurring in at least four of the nine disciplines), and discipline-specific (occurring in only one or two disciplines). Results indicate that these bundle types typically serve different functions. The cross-disciplinary bundles function as discourse frames providing sign posts for readers, while discipline-specific bundles are often content-specific. Analyses also identified bundles with variable slots that could be completed with various fillers. The chapter concludes with ideas and examples for teaching lexical bundles that can enhance teaching and promote fluency for learners.;;While corpus studies on academic writing have improved instructional materials for writing in the hard sciences, humanities and social sciences world language writing pedagogy remains open to development. In the interest of data-driven Spanish for academic purposes curricula, using English and Spanish corpora of psychology, history and literary criticism articles, we analyzed nouns occurring in sentence-subject position in 100 randomly-sampled sentences. In both languages, psychology had significantly more epistemic, and fewer phenomenal, sentence subject nouns than the other two fields. We extended our analysis to lexical bundles containing Spanish-English equivalent noun phrases in sentence-subject position which occurred significantly more often in psychology in both languages. The results are discussed in terms of scientific and humanities writing pedagogy for both languages.;;This chapter analyzes the use of attributive adjectives as nominal pre-modifiers in two corpora: CorAChem (Corpus of Articles in Chemistry) and CorAAL (Corpus of Articles in Applied Linguistics) composed of 150 articles each. Our aim is to understand the use of pre-modifying adjectives in noun phrases (NPs), considering which adjectives are used as well as the NP size and frequency across disciplines. The results show that both corpora carry more classifiers than descriptors. Nevertheless, each discipline favors the use of adjectives with specific functions. Both corpora use long premodifying sequences, but CorAChem carries a greater number of longer sequences. The use of long NPs may affect understanding due to the interrelation among the NP constituents as shown in CorAChem.;;Over the past decade, engineering has been one of the most oft-chosen fields of study for domestic and international students at US colleges and universities (IIE 2017; NCES 2018). As a result, the number of students in undergraduate engineering courses has steadily grown, leading to increased attention devoted to the written discourse encountered in engineering courses. The present study investigates the linguistic overlap between published textbooks used in lower-division undergraduate engineering courses by focusing on the analysis of multi-word sequences – 5-word phrase-frames – commonly found in five engineering disciplines. Overall, it was found that, while frequency distribution and structural characteristics of the identified phrase-frames were consistent across the five corpora, there were dissimilarities in the discourse functions performed by these patterns.;;This exploratory study investigates modal verbs as stance features in the Physical Sciences sub-corpus of the Michigan Corpus of Upper-Level Student Papers (MICUSP). The study adopts Biber’s (2006) stance framework, including modal verbs in three categories: possibility, necessity, and prediction. Focusing on one feature within one sub-corpus afforded closer consideration of the units of analysis (discipline, level of study, nativeness, and register) that could contribute to variation in the use of modals as stance features. The findings show that possibility and prediction modals are the most common in the Physical Sciences sub-corpus. The study provides a description of student academic writing and considers future research directions and pedagogical implications of stance in Physical Science disciplines, student levels, nativeness, and registers.;;This chapter examines the use of formulaic phrase frames (p-frames) in the rhetorical construction of Applied Linguistics Conference Abstracts (CAs). We consider both the distribution of p-frames across rhetorical moves/steps and the strength of the association of p-frames with individual moves/steps in 625 accepted AAAL 2017 CAs, all freely available on the organization’s website. We manually tagged the corpus for rhetorical moves/steps and identified p-frames of 5 and 6 words (29 per million threshold) using kfNgram (Fletcher 2007). Results highlight that p-frames play a prominent role in the realization of CA writers’ rhetorical goals both in the prevalence of p-frame use (roughly 1.7 instances per text in our data) and in the close relationship that many p-frames exhibit with particular rhetorical functions.;;This chapter presents a new multi-dimensional analysis of stand-alone literature reviews (e.g., Annual Review of Medicine ). Using the methods of Biber (1988), I analyzed 417 stand-alone literature reviews. A factor analysis of 72 grammatical features was conducted and a six-factor solution chosen. The dimensions were interpreted as Human vs. Technical/Academic Focus, Questioning/ Interpreting vs. Knowledge-Conferring, Expression of Stance, Author/Discourse Community vs. Topic Focus, Abstract vs. Concrete Focus, and Methodological Concerns vs. Description. Texts were then divided into four groups: discipline, time period, review type, and the presence or absence of methods. Factorial ANOVAs found significant differences for discipline and time period for many dimensions, whereas significant differences for presence of methods and review type were limited to one dimension each.;;This chapter discusses the identification of the major sets of interrelated collocations in academic writing across different disciplines, or dimensions of collocation. A corpus of textbooks and research articles from Social, Behavioral and Economic Sciences, containing 10.6 million words across 230 texts was analyzed using a collocational multi-dimensional analysis framework (Berber Sardinha 2017). Results show four dimensions of collocation for Social, Behavioral and Economic Sciences, namely: (1) Human Nature, Culture and Research Methods vs. Economics; (2) Human Evolution and Society; (3) Business and Finance; and (4) Statistics. The chapter highlights a methodology for identifying collocational patterns through multi-dimensional analysis, and contributes to a better understanding of academic writing in Social, Behavioral and Economic Sciences.