CHAPTER 2 LITERATURE REVIEW
2.6 Psycholinguistic Experiments
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Thus, another approach to explore the language preference of users will be demonstrated in the next section.
2.6 Psycholinguistic Experiments
In order to examine what factors may affect the language use, some scholars (Cronin & Jreisat, 1995; Thomson, Murachver, & Green, 2001) proposed that psycholinguistic experiments can be applied to identify the language preference of users.
Cronin and Jreisat (1995) explored whether teacher’s modeling, grade, gender, and attitude toward women would affect thigh school students’ use of nonsexist language or sexist language. In the study, sexist language was defined as the use of gender specific pronoun (e.g., he) to refer both sexes. Three types of questionnaires were conducted as follow: (a) The Sexist Language Detector (SLD) used to explore the participants’ use of sexist language by asking them to write down the solutions to ethical dilemma, (b) Personal Attributes Questionnaire (PAQ), used to predict the participant’s personality characteristics (e.g., self-assertive, independent, interpersonal-oriented, and expressive qualities), and (c) Attitudes toward Women Scale (ATW) used to understand the relation between personal characters and their uses of sexist language. The results indicated that teachers’ ‘modeling’ of nonsexist examples had a positive effect on the participants’ performances on using the nonsexist language. Furthermore, the grade of students and the gender were also related to the nonsexist language use. Freshman students used more nonsexist language than senior ones. The nonsexist language was more frequently used by the female participants, compared to the male participants.
Cronin and Jreisat (1995, p. 828) suggested that one of the possible reasons was that the male participants used less nonsexist language may due to their resistance to change.
Teachers may need different strategies to teach males and females to use the nonsexist language.
In another study, Thomson, Murachver, & Green (2001) aimed to examine how
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
women and men reacted and accommodated to ‘gender-preferential language’ in an e-mail context. The language styles of the genders were based on the features found in a previous study (Thomson & Murachver, 2001). For example, females tended to use
“words of questions, self-derogatory comments, compliments, apologies and subjective conjunctions. Males were more likely to convey opinions, make insults and to write more” (Thomson & Murachver, 2001, p. 200). There were two experiments in Thomson’s (2001) study. In the first experiment, the participants were asked to write e-mail letters to two assigned net pals who were actually one of the experimenters. The gender-preferential language used by the two net pals matched with their genders. The results showed that the participants’ language use were affected by their net pals’
language styles, rather than the genders of themselves. In order to examine to what extent that the participants’ language use was predicted by their net pals’ language style, the ‘language style of net pals’ and ‘gender labels’ were independently manipulated in the second experiment. The same group of participants were asked to write messages to one net pal only. The net pal used either fepreferential language style or male-preferential language style and either a male label (e.g., John) or a female label (e.g., Kate). There were four conditions in the study (i.e., language styles ⅹ gender labels=2
ⅹ 2=4). According to the second experiment, the researchers suggested that the language styles of the net pals had the greatest impact on the participants’ language use.
Overall, from the study, one could know that each person had the ability to use a range of language styles depending on whom he or she was talking to. “The gendered language was related to the communicative context. But it didn’t mean that “the gendered language is an invisible, inherent feature of a person” (Thomson, Murachver,
& Green, 2001, p. 174).
Overall, from the overview of the prior related studies in this chapter, the distinction between pretty and charming should be investigated from the corpus
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
analysis. It can help us understand the different use of the two adjectives from the perspective of English native speakers, such as different behaviors shown in collocation, syntax, and register. On the other hand, the relation between the native language speaker’s use of the two adjectives and the EFL learner’s difficulties of learning English could be elaborated profoundly through an integrated approach with a corpus analysis and psycholinguistic experiment. Therefore, in the following
chapters, the thesis will focus on the two issues: the native speaker’s use of pretty and charming (Chapter Three) and identification of the learning difficulties of the pair of near-synonyms (Chapter Four).
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
CHAPTER 3 CORPUS ANALYSIS
Based on the overview of the previous studies on discourse of gender, few studies could be found integrating corpus analysis and psycholinguistic experiments. We realized that adopting both research methods could be helpful for us to have a better understanding of how language is used to represent male and female. Therefore, in this chapter, the first part will focus on the corpus analysis, the methodology for the following three aspects will be introduced (a) to establish a coding scheme for data classification, (b) to analyze the semantic distribution of pretty and charming; and (c) to investigate the collocates of the two adjectives. In the following part of this chapter, the results of corpus analysis will be illustrated.
3.1 Methodology of the Corpus Study
3.1.1 The Corpus—the Corpus of Contemporary American English (COCA) To investigate the similarities and differences of pretty and charming, the Corpus of Contemporary American English (COCA) was adopted as a source data in the study.
The corpus contains more than 520 million words of texts from 1990 to 2015, including spoken, fiction, magazines, newspapers and academic texts. The distribution of texts is equally divided (COCA). The overall frequency patterns of pretty and charming in the adjective forms in the COCA were 17136 and 6171 instances respectively. For the present study, a total of 500 instances of the concordance results were randomly retrieved for each adjective. To identify adjective uses of each target word, a search of part-of-speech tagging was conducted. The results contained some mistagged instances since pretty can also function as adverb, for example, Water is pretty cold up there (COCA), pretty here functioning as adverb rather than adjective. It was required to exclude these mistagged instances by manual annotation. Moreover, if an instance has
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
more than one target word, each one will be counted. For instance, in (3-1), the target word, pretty, appears in the instance two times.
(3-1) He wasn't about to give a job to something fresh off the street with --so far as he could tell – nothing on the ball but looks. Pretty? The club had pretty faces to burn. Gorgeous? That was another story.
3.1.2 The Establishment of a Coding Scheme
In order to detect the differences between pretty and charming, a coding scheme was established based on (a) word senses from three online dictionaries, (b) the results of most frequently appearing nouns modified by the two adjectives, and (c) an adaption of category scheme. For example (3-1), the first pretty was used to describe the job applicant’s appearance which was labeled as [Person], [Outer beauty] and [Beauty]; the second pretty describes the club’s appearance or nature, labeling as [Object] [Concrete], and [Place]. The definitions of types of classification and examples are based on Haily and Jung’s study (2015).
Word senses of pretty and charming from dictionaries. Three online dictionaries
were used to examine the word senses of the two target words. The three dictionaries are Oxford English Online Dictionary, Merriam-Webster Dictionary, and Collins’
COBUILD Advanced Learners’ English Dictionary. These dictionaries are freely accessible and frequently updated, providing users with information of the latest words or phrases from both British and American usages of English. The overall definitions of the two adjectives are shown in Table 3-1 and Table 3-2. It appears that pretty and charming have one shared sense— both of them can be used to describe someone or something as ‘pleasant’ or ‘attractive’. The nuance between them may be that charming is a stronger expression to modify a person or thing because its definitions from three dictionaries include intensifiers (the underlined words in Table 3-2), such as ‘very’ and
‧
Table 3- 1 Definitions of ‘Pretty’ from Three Dictionaries pretty
(a) pleasing by delicacy or grace
(b) having conventionally accepted elements of beauty
(c) appearing or sounding pleasant or nice but
Sense 3 (a) miserable, terrible
(b) chiefly Scotland :stout
Sense 4 moderately
large: considerable
Sense 5 easy to enjoy: pleasant —
usually used in negative constructions
‧
Table 3- 2 Definitions of ‘Charming’ from Three Dictionaries charming
a charming restaurant
has such charming
However, there are some differences between pretty and charming, when they are used to describe people. The former one tends to describe someone’s appearance (you have no idea how pretty you look a woman of your age) and the latter one tends to describe someone’s manner or behavior (he was a charming, affectionate colleague).
In addition, when describing someone as pretty, all definitions provided by the three dictionaries used ‘delicacy’ or its adjective form (i.e., ‘delicate’) to define the meaning. To further understand the meaning of ‘delicacy’, a search in WordNet 3.1 was conducted. The result shows that delicacy refers to (a) ‘smallness of stature’, (b) ‘lack
‧
of physical strength’, and (c) ‘lightness in movement or manner’. The examples below are derived from the COCA. As for (3-5), pretty is used to depict someone’s appearance which corresponds to sense 2 in the Merriam-Webster Dictionary (i.e., ‘having conventionally accepted elements of beauty’). Here, Table 3-3, all the three dimensions are exemplified as (3-2) – (3-4).
Table 3- 3 The Senses of ‘Pretty’ in the Dictionary
Types Definitions Examples
(a) Slightness Someone is thin or small in stature.
(3-2) Sometimes Korean girls get part of their leg muscles removed, to be pretty.
(b) Fragility Someone lacks strength
(3-3) …someone who was very, very angry with this particular pretty girl plunged an ordinary dinner knife into her pretty chest.
(c) Airiness someone is light and graceful in
movement
(3-4) Her inability to sit still, and her grace that made you catch your breath. Libbie was said to be pretty…
Apart from the word senses shown in the Table 3-1 and Table 3-1, both pretty and charming also have a discourse function of expressing disapproving, a negative use.
For pretty, three negative meanings are found in the dictionaries in this study : (a)
‘miserable, terrible’, (b) something is not ‘easy to enjoy’, and (c) to be ‘used ironically to express annoyance or displeasure’.
(3-6) Negative uses of pretty (Merriam-Webster Dictionary) (a) ‘miserable, terrible’
e.g., You try not to go into the real ugly cry on TV, because this is the thing. If
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
you're watching somebody cry and you're kind of with them as long as they're kind of crying pretty (COCA).
(b) something is not ‘easy to enjoy’ (Merriam-Webster Dictionary) e.g., I already know how my welcome is going to be. It's not going to be too pretty, ‘he said’ (COCA).
(c) ‘used ironically to express annoyance or displeasure’ (Oxford Dictionary) e.g., he led me a pretty dance (Oxford Dictionary)
As for charming, negative uses were not found in these three dictionaries. However, we will show some examples that are found in the corpus data.
In summary, based on the definition of words from three dictionaries mentioned above, the main meaning shared by pretty and charming is ‘pleasant or attractive’, and both of them can be used to describe people, things or places. Furthermore, the senses shown in the dictionaries also imply that the two adjectives have both positive and negative meanings.
Collocates of pretty and charming in the COCA. To establish a coding scheme
for data classification, R1 collocates were searched to identify types of nouns which the two adjectives modified. Two types of queries were used for searching the R1 modified nouns; that is, frequency and MI score. As Liu’s study (2010, p. 63) pointed out that adopting both the two queries may complement each other in better identifying the typical nouns modified by the adjectives. In order to establish the representativeness of the most frequent nouns measured by MI scores, the nouns that have fewer than 35 tokens of co-occurrence with pretty were excluded. Given that the lower overall frequency of charming, the criteria of selection were lowered down to 10. Thus, the nouns which have fewer than 10 tokens of co-occurrence with charming were excluded.
The top 20 nouns modified by the two adjectives were recorded. A preliminary
‧
classification of the types of modified nouns was conducted and the results were shown in the following Table.
Table 3- 4 Types of Nouns Modified by ‘Pretty’ and ‘Charming’
Concrete Abstract Dual
pretty girl(s),woman (women), lady, boy(s), baby, face(s), eyes, clothes, dress(es), flowers, horses, penny, smile
name picture(s), sight,
thing(s) charming man (men), guy, person, people, smile
woman, fellow, child, wife, girl, hostess, lady, grin, town, place, village, cottage, hotel, book
story, personality, scene, accent
way
There are three major categories for the modified nouns: (1) concrete nouns, i.e., a noun that refers to a material object (Collins Dictionary), (2) abstract nouns, i.e., a noun that refers to a thing that does not exist as a material object (Collins Dictionary), and (3) dual nouns, i.e., a noun can be both concrete and abstract in different contexts.
As shown in Table 3-4, the adjectives are frequently collocated with concrete nouns and words related to people (e.g., girl, woman, etc.). Moreover, when describing someone’s appearance, it seems that facial features or expressions (e.g., face and smile in examples below) are essential elements to show their attractions.
(3-7) Mainly because she had the advantage of knowing that it took more than big breasts, a slim waist, onion booty and a pretty face to grab and hold his attention.
(3-8) It's my favorite city. He gives him his most charming smile.….
In sum, based on the classification of the top 20 noun collocates of the two adjectives, it seems that this result provides a different perspective on the understanding of the two adjectives compared to the observation of the word senses in dictionaries.
From the results of the analysis of collocate, pretty and charming are more commonly
‧
Whereas, for word senses, pretty is frequently used to describe someone’s appearance and it is especially used for describing females. On the other hand, charming is used to describe someone’s character or attitude. Based on this finding, the issue of sources of beauty (i.e., inner or outer beauty) and the gender representation in uses of the two adjectives are still pending. Since the word senses and collocational information can only provide some semantic information of the target words, a further investigation is needed by using annotating the corpus data. In the following part, we will introduce the development of a coding scheme for concordance lines analysis.
A coding scheme. Based on the word senses and the results of most frequently
appearing nouns modified by pretty and charming, a coding scheme was established (Table 3-5 and Table 3-6) for further analyzing semantic distribution. In the coding scheme, there are two categories modified by pretty and charming: person and object.
Table 3- 5 Coding Scheme of Person
Modified nouns Types
As shown in Table 3-5, someone’s attractiveness can be seen from one’s ‘outer beauty’ as well as ‘inner beauty’. Moreover, the attractiveness of ‘outer beauty’ was further divided into four main types based on the senses of pretty— ‘slightness’,
‘fragility’, ‘airiness’ and ‘beauty’ provided by the dictionary definitions. The ‘inner
beauty’ referred to the ‘character’ of a person.
In terms of the attractiveness of objects, all the modified nouns were firstly divided into two types: ‘concrete’ and ‘abstract’. In this study, concrete words were defined as
‧
the words which can be touched or sensed in any other ways (Khokhlova, 2014).
Table 3- 6 Coding Scheme of Object
Types Modified nouns Examples
Object 1. Concrete 2. Abstract
(a) Artifact picture, story, dress (b) Place barn, garden, house (c) Sound crooning, music (d) Nature beach, tree, roses (e) General nouns things, stuff, one (f) Others sitting pretty
Then, in each type of modified nouns, they were further arranged in six categories.
These six categories were established by the results of classification of R1 noun collocates (see Chapter 3, Table 3-12) and the adaption of category scheme from Haily and Jung’s study (2015, p. 140).
The first category is ‘artifact’, referring to something created by people (e.g. dress, painting, ideas, and joke). The second and third are a ‘place’ and ‘sound’ which refer to
‘any point, building, area, town, or country’ and ‘something that you can hear’
respectively. ‘Nature’ refers to ‘all the animals, plants, and other things in the world that are not made by people’ (Collins Dictionary). For the last two types of modified nouns— ‘general nouns’ and ‘others’, the former one refers to nouns with referential functions, such as one, thing, and stuff (e.g. …countries have contributed to the situation along the border, which is indeed not a pretty one). The latter one contains those nouns that could not be assigned to any of the five types and the idiomatic usages (e.g. sitting pretty, not a pretty sight).
3.2 Results of the Corpus Study
This section presents the results of corpus analysis of the two adjectives—pretty and charming. For the present study, a total of 500 instances of the concordance results were randomly retrieved for each adjective from the COCA. After the extraction
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
process introduced previously, 907 out of 1000 instances were kept for further analyses (i.e., 464 instances for pretty and 443 for charming). The results will be presented in the following parts: the overall semantic distribution, the distribution of sex roles of modified people, collocational information, and a comparison of the two adjectives in different genres.
3.2.1 Semantic Distribution
The sense analysis of pretty and charming can be displayed in two categories based on the modified nouns— ‘person’ and ‘object’. As Table 3-7 shows that pretty and charming were commonly used to describe people or things as attractive. Moreover, the two adjectives showed a slight preference toward describe ‘person’ with about 52% for pretty and 53% for charming.
Table 3- 7 The Distributions of Modified Nouns for ‘Pretty’ and ‘Charming’
pretty charming
Modified Nouns Freq. % Freq. %
Person 239 51.51% 234 52.82%
Object 218 46.98% 205 46.28%
Animal 7 1.51% 4 0.90%
Total 464 100.00% 443 100.00%
Modified nouns—person. Regarding modified nouns related to person, Table 3-8
present the percentages of inner and outer beauty described by pretty and charming.
The results indicate that pretty seems to be more commonly used to modify someone’s physical appearance tagged as ‘beauty’ (90.79%), whereas charming is more commonly used to describe someone’s character (82.48%). Moreover, as previously mentioned, all the dictionaries adopted in the study used the word ‘delicacy’ or its adjective form (i.e., delicate) to define the meaning of pretty. Three types of ‘delicacy’ were established for analyzing, that is, ‘slightness’, ‘fragility’, and ‘airiness’. The results show that only a
‧
few instances displayed the meaning of ‘delicacy’ with 6.7% of the total instances. In these examples, we found that pretty was likely to collocate with other adjectives having similar senses of delicate, such as little, thin and tiny. For instances, in example (3-9), little and petite corresponds to the sense of ‘slightness’. The adjectives were used to describe someone as thin or small in stature. Furthermore, some of the instances displayed that there was a cause and effect relationship between the behavior of the person described in the sentence and the beauty he or she admired. In example (3-10), Korean girls wanted to be pretty. To attain the goal, they tried to get part of their muscles removed. It seems that what the girls pursued was to be thin, and to have a good body shape (tagged as ‘slightness’).
(3-9) She was a pretty little thing, petite and curvy, with a quick wit and generous nature…
(3-10) “Sometimes Korean girls get part of their leg muscles removed, to be pretty.
You can get your hand half around their calves, “said one.
You can get your hand half around their calves, “said one.