國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
99
CHAPTER 5 DISCUSSION
The issue of distinguishing the meanings of in and on through semantic feature analysis in combination with sense analysis has been the focus of this thesis. This chapter will discuss these findings with the support of related theories, organized as the following. First, a summary of the major findings in Chapter Four is provided in 5.1. Then, in Section 5.2, research findings are discussed against the research questions based on the results obtained herein.
5.1 Major Findings
As the prepositions in and on have many overlapping yet different meanings, their variety in senses and usages is worth exploring. This study adopted the semantic feature analysis in combination with sense analysis in order to provide a systematic approach of analyzing in and on. This quantitative approach may help distinguish the semantic meanings of in and on through looking into the kinds of figure and ground nouns and their interaction in forming the prepositional constructions. With regard to the comparison between native speaker‘s and L2 learner‘s performance in using these two prepositions, the result was also run to compare the similarities and differences of semantic features and senses identified in these two sets of data. This method is different from previous related research as previous studies tend to focus on exploring the senses of a particular preposition or inspecting learners‘ performance in the use of a set of prepositions in some elicitation experiments. However, in this study, we investigated both the senses and learners‘ performance in using two prepositions at
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
100
the same time, with the aim to provide a more comprehensive result toward this issue.
In the first section of Chapter Four, we firstly looked into the result of sense analysis, from which we attempted to investigate if there is any difference between native speakers‘ and L2 learners‘ use of the prepositions in and on. The result showed that, for preposition in, most of the instances could be grouped into the ―inclusion, location, or position within limits‖ category, which is the closest meaning to whose proto-scene in the semantic network of in (Evans & Tyler, 2004). The second-most frequent one is the sense ―used as a function word to indicate limitation, qualification, or circumstance‖, which is about two times higher in BNC than NCCU. Only slight variations between BNC and NCCU can be found in other categories, implying that native speakers and learners‘ understanding and command of the senses and usages do not differ greatly in the construction of in. For the result of preposition on, in senses that refer to location, higher frequency could be found in NCCU. More metaphorical constructions of on were identified in the data of BNC, which implied that learners‘
command of on is more restricted and thus learners may use the locational sense of on extensively in the data. This may also result from the complexity in the senses of on, which vary diversely from literal to metaphorical extended meanings, engendering the avoidance strategy in using unfamiliar senses and constructions of on. With regard to semantic feature analysis, the distribution of semantic feature is compared between the literal and metaphorical constructions of in and on respectively, as summarized in Table 5.1.
‧
Table 5.1 List of Statistically Prominent Features in BNC
Literal Metaphorical
Figure Ground Figure Ground
In
Figure Ground Figure Ground
On
For figure nouns in literal constructions, in the case of in, the figure nouns tend to be animate (and human) entity which is active in moving. Slight difference was found with on, in that animate feature is no longer prominent in the data; instead, it tends to go with entities that are concrete, mobile, countable, and solid (e.g. handbag, electrocandles). As for features in ground nouns of literal constructions, mostly concrete nouns are prominent in the data of in, which refer to places or locations in the real world. Concrete feature was identified as a prominent feature in the data of on (e.g. on page 8, on Mars); however, adding the meaning of support, that is, the
ground acting as a supporting surface for the figure to be situated, ground nouns are more mobile and solid as well (e.g. on the board, on the seat).
The results in the metaphorical expressions are more complex since greater complexity in these extended senses can be found in both in and on. As the variation among features in senses is great, the distribution of semantic features cannot be
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
102
summed up by simply showing the tendency in all metaphorical instances, but has to be compared separately in those prominent senses. Therefore, the results in this part were also discussed based on the metaphorical prominent sense categories identified in the sense analysis. For figure nouns of in, [-animate] feature tends to appear more frequently in the data of the MEANS cluster. These [-animate] objects are usually concrete things which are closely related to human activity or for completing an action (e.g. in brown, in line). Features that are prominent in the data may vary from concrete to non-concrete in metaphorical expressions of in, as in the sense referring to
―something in vogue or season‖, the figure may be concrete nouns (e.g. clothing, blue) and may also be style or activity (e.g. action, fashion), etc. Similar cases in the variation of prominent semantic features can also be observed in the data of the metaphorical on constructions. For example, examining the results from those
metaphorical prominent senses, we found that the nouns in the SUPPORT cluster (e.g.
tickets, money) are more concrete and solid, but they may also contain activity feature (e.g. choice, reliance). For ground nouns in the metaphorical in constructions,
measure and activity features were found to appear frequently in the data as in the LOCATION cluster, as there is a ―state‖ sense which shows the condition of an entity as well as an ―activity‖ sense indicating where the figure occurs. In contrast to the tendency to be [-animate] in in, in the metaphorical prominent senses of on, concrete, solid, and countable features tend to occur more frequently. Being similar to the grounds in literal constructions, the grounds in metaphorical constructions are usually real world objects acting as locations that constrain the occurrence of an activity.
From the findings in this part, we claim that there is indeed a tendency that certain type of noun goes with a particular expression with a preposition, and the prominent features within these nouns may vary from literal to different metaphorical senses of
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
103
that particular preposition.
In addition to the comparison between the semantic feature distributions of in and on, the comparison within each preposition group was also conducted between the native speaker and learner data, as summarized in Figure 5.1.
Figure 5.1 The Comparison of Prominent Semantic Features Between BNC and NCCU
As summarized in Figure 5.1, the content in the grids on each column shows the prominent type of nouns in that category. For the cases of literal in-constructions, more human names or references to human beings were found in the figure nouns in the learner data, while greater variety can be seen in the figures of native speaker data.
As for the grounds, there is a tendency of fewer solid grounds in learner data as regions or country names account for a high frequency in the data, but this tendency was not observed in the data from BNC. For metaphorical in-construction, no great difference in figure nouns was found between the data from BNC and NCCU, but in
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
104
ground nouns, activity and solid features were identified with higher frequency in native speaker data, which are not the case in NCCU. With regard to on prepositional phrases produced by the learner, in literal constructions, human names or references to people were used more frequently in figure nouns, and higher number of mobile, countable, and solid features were used in ground nouns. In metaphorical data of on, L2 learners were still inclined to use animate and human figure nouns while native speakers used more solid figure nouns. As for ground nouns in metaphorical on constructions, there are also higher number of solid nouns found in native speaker data. However, for the metaphorical set of data, due to the complexity in sense, the results may vary within different metaphorical prominent senses. From these results we can observe that though in sense analysis, no great difference was found between the distribution of senses in BNC and NCCU, inconsistency in the distribution of the types of nouns can still be seen in semantic feature analysis. The more various the types of nouns are involved in, the less prominent features may be found as these varieties of nouns may not be able to be described using the same sets of semantic features. If a set of nouns vary diversely, it is less likely to be described using many different semantic features. Therefore, we can conclude that learners‘ use of figure and ground nouns in constructing prepositional expressions is more restricted to particular nouns.
In the last section of Chapter 4, the errors identified in the learner data were discussed respectively according to some common error types, including substitution of other prepositions, redundancy, and misuse of a particular preposition. In fact, not many errors were found in the data produced by the learners, so each error type in both in and on was discussed to probe into the possible causes of the error.
In next section, these findings will be discussed to investigate the formation of
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
105
prepositional phrases of particular senses. Moreover, in order to facilitate the learning of English prepositions, these findings are also extended to probe into the learners‘
tendency and problems in producing prepositional phrases.