Siaw-Fong Chung
3.0 Methodology and Results
This study examines the use of the words ENGINE, INFORMATION and DATA in British National Corpus (BNC) (100 million words) and compares them to their uses in 8,368 dissertation abstracts (1,894,299 words) produced by graduate students from the Management and Information Science (MIS) departments from fifty-five universities in Taiwan.1 Table 1 below shows the number of hits found with each of these words and their derived forms.
Their frequency per thousand and per million words is shown in the last two rows.
1 Capitalized words are used to refer to lemma or words representing all their derived forms (-s, -ed, etc.)
Table 1: The Number of Instances for ENGINE, INFORMATION and DATA used in Dissertation Writing vs.
BNC
Word Theses (%)
BNC Word Theses BNC Word Theses BNC
Engine 280
From Table 1, we can see that ENGINE
appears most often (43.88%) as engineering in the
theses, whereas in a general corpus like BNC, it occurs most often (41.94%) as engine. ForINFORMATION and DATA, the results do not differ greatly between the theses corpus and BNC.
For all three words, the theses corpus shows to return more instances of ENGINE, INFORMATION
and DATA when calculated in per thousand and per million words. In order to see further how these three keywords can be used in verb-particle constructions, all constructions comprising a verb plus a particle when used with ENGINE, INFORMATION and DATA are extracted through BNCWeb, a platform used to access BNC (Hoffmann, Evert, Smith et al., 2008). The results display that in BNC, based on native-speakers’ data, engine and its lemma appear 48 times after a verb-particle construction, indicating a small percentage of 0.42 from the total 11,518 instances of ENGINE. For INFORMATION, its verb-particle constructions return 735 instances,
of verb-particle constructions from its total 18,097 instances. These percentages show that verb-particle constructions are rare when they are used with ENGINE, INFORMATION and DATA. In order to observe what kinds of verb-particle constructions are found with the three keywords, the following Table 2 from BNC is presented. Table 2 displays the top twenty results for all occurrences of verbs, followed by any particles (including prepositions), which precede the three keywords of interest.2
Table 2: Instances of Verb+Particle+Keyword in BNC
Lexical Items Freq. % Lexical Items Freq. % Lexical Items Freq. % tinkering with
engines 2 4.17% based on information 50 6.8% based on data 27 9.47%
comes to engines 1 2.08% asking for information 26 3.54% used for data 10 3.51%
quit with engine 1 2.08% pass on information 17 2.31% derived from data 9 3.16%
concentrated on
engines 1 2.08% according to information 15 2.04% According to data 9 3.16%
depending on engine 1 2.08% asked for information 14 1.9% follow up data 6 2.11%
ran into engine 1 2.08% passing on information 14 1.9% supported by data 5 1.75%
used by engines 1 2.08% appealed for information 12 1.63% draws on data 4 1.4%
say about engine 1 2.08% ask for information 12 1.63% treated as data 4 1.4%
Switch off engine 1 2.08% looking for information 10 1.36% compiled from data 3 1.05%
depends on engine 1 2.08% packed with information 10 1.36% combined with data 3 1.05%
take over engine 1 2.08% offered for information 9 1.22% compared with data 3 1.05%
surrounded by
engine 1 2.08% appealing for information 9 1.22% relate to data 3 1.05%
been in engine 1 2.08% provided with
information 8 1.09% known as data 3 1.05%
do on engine 1 2.08% searching for information 8 1.09% used as data 3 1.05%
looking after engines 1 2.08% search for information 7 0.95% taken from Data 3 1.05%
sitting with engine 1 2.08% look for information 7 0.95% qualify as data 3 1.05%
2 Table 1 shows all verb+particle constructions which include all occurrences of a particle after a verb. A automatic extraction of them from corpora may return noise such as the unwanted ones (e.g., been in engine).
engine information
knew about engines 1 2.08% finding out information 6 0.82% spent in data 2 0.7%
are in engine 1 2.08% FOLLOW UP
INFORMATION 6 0.82% associated with
data 2 0.7%
used as engine 1 2.08% refer to information 6 0.82% picks out data 2 0.7%
finished with
engines 1 2.08% seek out information 5 0.68% do with data 2 0.7%
From all the uses in Table 2, only a few (shaded) are found to be phrasal verbs. Most of the uses are prepositional verbs, i.e., the combination of the verb and the preposition is decomposeable. When we investigate the theses, only a less than 5% of verb-particle constructions are found. Most of them are highly collocated with based on, according to,
search for, etc., which are mostly prepositional verbs.
In addition, for the theses corpus, we also found that engine (including engineer and
engineering) and information are often used to differentiate different types of nominals such
that in differentiating types of engineers (software engineers, system engineers andknowledge engineers) and types of information (enterprise information, medical information).
In contrast, data is more often used when a particular action carried out towards the data (e.g.,
collect data, data mining, data exchange). When we examine data versus information in a
web-based language collected in a corpus called UKWac 1.0 (Bailey & Thompson, 2006), the results further confirm our observation. For instance, only data can be used with observed orobservational, while only information can be used with travel, tourist, health, etc. Our
findings show that some terminologies in engineering are preferred to refer to the types of things while some terms are used to refer to the actions (even through these terms might be nouns).4.0 Conclusion
With one of the aims to investigate verb-particle constructions in English through using corpora data, this paper has compared a specialized corpus constituted by academic dissertation abstracts written by EFL students of Taiwan to a native-speaker corpus. The results discussed in this work serve as a pilot study and the topic deserves greater discussion.
The findings herein will become important to advisory work when directing students’
dissertation writing. Despite the fact that Jackendoff (1997) has claimed phrasal lexicon is about the same size as the simplex (i.e., single-word) lexicon used by an average native speaker of English, this study found that verb-particle constructions are probably used less in learners’ writing, when compared to native speakers’. Therefore, it is worth pondering
expressions, of which verb-particle constructions are one type of them, has been said to be crucial to becoming native-speaker like in language proficiency. This study shows that, linguistic use of words, when taught to students from the information science background, will help raise the students’ awareness in writing. This study, therefore, calls for the attention towards specific use of lexis in the writing process of non-humanity students.
References
Anderson, J., & Poole, M. 1994. Thesis and Assignment Writing (2nd ed.). Brisbane: John Wiley.
Bailey, S. & Thompson, D. 2006. “UKWAC: Building the UK's First Public Web Archive.”