Final

Introduction

Introductions play an important role in academic articles because it will give the audience a clear clue about the topic and structure of the paper. Therefore, the question "What structure should an introduction follow" has drawn much attention by linguists. Corpus had been investigated in a variety of disciplines and some models were proposed. For example, Swales put forward CARS (Create A Research Space) model and revised it in 2004, breaking down introductions into three moves: establishing the area of research, establishing the gap in knowledge, and filling this gap. This model was frequently examined and researchers found out that most introductions in all kinds of discipline follow it well.

After CARS model had been revised and accepted as useful, it was widely applied to analysis in introductions of research articles. Especially in recent years, several papers were proposed each year discussing about the rhetorical element, crosslinguistic expression or other aspects of introductions. However, during such a long time only a few researchers paid close attention to the structural difference between introductions and other elements of research articles when applying CARS model (Riley, 1991; Samraj, 2005; Pho, 2008), while these other elements such as abstracts and conclusions are structurally similar to introductions. For example, abstracts also require a brief generalization of the context of the article, mentioning the research process and result which is similar to the Move 3 of introductions regarding CARS model. Therefore, question arises that "Do other parts of research article follow CARS model as introductions and why are they structurally different in a certain discipline?" The goal of this research is to find what is the structural difference between introductions and abstracts, and what factor causes such difference. The result will help authors to better understand how to write introduction.

Methodology

Introduction and abstracts are similar. They both serve as a summary, indicating the reason, the topic and maybe to some extent the result of the article. In order to compare them, a corpus of 12 research articles is adopted. The articles should be chosen in a variety of sources, from conference papers to students' dissertation in biomedical engineering field, with different length as a possible factor affecting the structure of introductions and abstracts. In this paper, the sources are journals, conferences, and dissertations. Each has 4 articles, written in recent years. Another possible factor is the purpose of the article, including to create a literature review of the area, to implement an algorithm, or to present a experimental finding. Such differences in purposes are also considered to be taken into account during the setup of the corpus.

Since CARS model is reliable and widely used, each introduction is analyzed under this model to see which moves it has. And then, to contrast with the introduction, the corresponding abstract is examined by CARS model. The research focus is counted as move one or two, while the methods used and findings are move three. The length of each move is also examined in a standard that moves less than 1/4 of the section are counted as "short", and those longer than 1/2 are counted as "long"; move length in between is "medium". Afterward contrast is made to see whether they are coherent with introductions, and why they differ if the two parts have the same structure.

Result

It turns out that most abstracts have similar structure of the corresponding introduction. The portion of the length of each move is similar in both parts. If a move in the abstract is "long", then the one in the introduction is also "long" or at least "medium". And if one is short, the corresponding one is often short or even absent. 7 out of 11 articles follow this pattern. For example, in LV Segmentation and Motion Analysis from 4D Cardiac Images (Yun Zhu, 2010), move one of the abstract is "medium" whereas that of the introduction is "short"; move two of the abstract have only one sentence while that of the introduction has just a few; both move threes are long. Another example is A Model-based Self-adaptive Approach to Image Processing (Jim Nichols & Ted Bapty, 2004). This article lacks move one in both abstract and introduction, and move two and three are weighed relatively equal. Another point is if there is a short or no abstract, then the introduction is usually short, like in Spotlights in Biomedical Imaging (Andrew F. Laine, 2009), where there is no abstract and the introduction has only two sentences.

Concerning the purpose of the articles, there is not much characteristic revealed except that 4 out of 4 articles with purpose of developing an algorithm have long move three in abstracts. The fact is that 7 out of 11 abstracts has long move three but no obvious trend of introductions.

Discussion

From the above analysis of the result we can see that authors tend to write in the same structure of abstracts and introductions. Usually abstracts are the shorten version of an introduction and they are often about the conclusion.

 Since abstracts and introductions are similar in structure, linguists may consider to develop a general model for the two parts. 

Final -- discussion

From the above analysis of the result we can see that authors tend to write in the same structure of abstracts and introductions. Usually abstracts are the shorten version of an introduction and they are often about the conclusion.
 Since abstracts and introductions are similar in structure, linguists may consider to develop a general model for the two parts.