Data of the Study and Data Source

The data collection process of this study involved three steps. Firstly, blends were searched carefully by using the search engine in Google website in order to get a large amount of data. After thorough searching, it was founded seven websites which provide rich data of blends such as Oxford, Merriam-Webster, Collins Dictionary, Cambridge, Word Spy, Rice University Neologisms Database, and Urban Dictionary. These seven online websites and dictionaries were then used as the primary sources of data. Secondly, blends were filtered in order to separate blends from “normal” words or any words which were not related to blends. One of the online dictionaries, The Rice University Neologisms Database, provided an advance search which enabled the researcher to search blends by simply typing “blends blending” as the keyword in the Search option. The rest of online websites and dictionaries, which did not have an advanced search, needed more thorough searching. Thirdly, blends were sorted due to some considerations. This process involved several procedures. The first procedure was sorting the repeated blends which also appeared in the different websites. This process was essential to avoid multiple data of the same entity. Taken from many sources, it was possible that there were several repeated blends which needed sorting. Finally, any repeated blends appearing in the different websites were not included. The second procedure was sorting the data due to their frequency of occurrence in the online dictionaries. Blends used for this research were only included if they were listed in at least two online dictionaries. It aimed at making sure that blends used in this research were not restricted to limited people and really used by people in everyday life conversations or in specific domains. The third procedure was eliminating three-word blends. It was due to the fact that blends were not merely the result of the combination of two words. Rather, the combination of three words was also possible. In addition to the limited number, three-word blends were not included since the researcher wanted to focus on two-word blends.

2. Data Analysis

In answering four research questions, a series of procedures were conducted. Morphological analysis was employed to answer the first and second research questions. Therefore, the analysis focused on answering these questions in which they dealt with structure and word category of blends. In terms of answering the third research question, the analysis of prosodic morphology was conducted since it examined the phonological characteristics of blends. The semantic analysis was employed to examine the fourth research question dealing with meaning patterns in blends. The results of these four analyses were included in the table as what presented in the following Table 3.1. The table showed how the researcher investigated blends using morphological, phonological, and semantic analysis. The first column contained blends and the source words. The source words consisted of two words: one as the first source word and the other one as the second source word. Henceforth, they were named SW1 and SW2. Furthermore, the first column presented the structure on how to combine elements in blends. Ways to „read‟ the structure of blend in the first column were presented as follows. As seen in the example number 2 of Table 3.1, the blend biathlete written as biathlon x athlete consisted of biathlon as the first source word SW1 and athlete as the second source word SW2. Besides, segments in bold signified overlapped segments and segments on brackets on were omitted segments. Different format was shown in blend number 1. In autopathography autobiography x pathology, the underlined segments indicated segments which were deleted. The segments in italics as in pathology refer to segments which were inserted to the other source word. The second column showed types of combining patterns in blends. The combining patterns showed variations so that the researcher needed to name each pattern. The patterns were named as type A, B, C, D, E, F, OV0, OV1a, OV1b, OV2a, and OV2b. Type A, B, C, and D belonged to clipping blends, yet the way of combining blends of each type was different. Type A combined whole SW1 and last part of SW2. Type B included blends which combined first part of SW1 and whole SW2. Type C belonged to blends which combined first part of SW1 and last part of SW2. Blends which belonged to type D joined first part of both SW1 and SW2 together. Meanwhile, type E and F belonged to blends with infixation and interfixation, respectively. Blends were included into type E if there was an infix which was put in the middle of SW1, while blends of type F had an interfix which linked first and second source words. Type OV0, OV1a, OV1b, OV2a, and OV2b belonged to overlapping blends. These types of blends also varied. Type OV0 meant that there was no clipping at all, however there were overlapped segments. Other remaining blends were blends of type OV1a whole + last, OV1b first + whole, OV2a first + last, and OV2b first + first. The third column showed word category of blends. It showed that the elements inside the brackets were the word category of source words, while the