Editor

JLLT edited by Thomas Tinnefeld
Showing posts with label 81 Chung. Show all posts
Showing posts with label 81 Chung. Show all posts
Journal of Linguistics and Language Teaching
Volume 8 (2017) Issue 1


A Corpus-Based Approach to Distinguishing the
Near-Synonyms Listen and Hear


Siaw-Fong Chung (Taipei, Taiwan (R.O.C.))


Abstract

The present study aimed to compare the verbs listen (to) and hear based on lexical resources and corpora, including (a) WordNet, which contains sense frequency information taken from the Brown Corpus and The Red Badge of Courage; (b) the British National Corpus (BNC); and (c) a writing task for English learners that focused on the uses of listen (to) and hear. The two verbs were compared in terms of sense frequency distributions as well as collocational information. Similarities and differences between the uses of the verbs listen (to) and hear were also analyzed using Sketch Engine, a lexical resource that enables collocational patterns from the BNC to be displayed according to grammatical relations. In the writing task, it was found that, for both verbs, students focused on a certain meaning. In addition, the BNC showed different sense distributions compared with WordNet and the learner data, as well as more figurative meanings. Both learner data and WordNet were predominant in the use of literal meanings for both verbs. This study contributes to the practice of connecting corpus data to teaching and learning. 
Keywords: near-synonyms, listen (to), hear, corpus, collocation, learner



1 Introduction

It has commonly been found that learners are often confused by the close meanings of near-synonymous words. Taylor (2003), for example, defines synonymy as
a single meaning [which] is symbolized by two or more distinct phonological forms” (Taylor 2003: 246)
indicating that synonyms are words that share a similar meaning. Nevertheless, many (e.g. Lyons 1968, Taylor 2003) have noted that perfect synonyms are very infrequent and that most synonyms are near-synonyms:
[P]erfect synonymy is vanishingly rare, methodologically proscribed, or a logical impossibility, what we frequently do encounter are pairs of words that are ‘‘near’’ synonyms” (Taylor 2003: 265)
In other words, two closely related words are seldom used in exactly the same way. Moreover, near-synonyms usually differ in a subtle way, as Bolinger (1977 noted,
if two ways of saying something differ in their words or their arrangement they will also differ in meaning.’ (Bolinger 1977: 1)
With regard to near-synonyms, some studies have focused on the semantic distinctions between the compared words. With the help of computer-aided technology, studies on near-synonyms can now be carried out, based on abundant computer-generated data. For example, in a corpus-based study on the English verbs start and begin, Biber et al. (1998) discovered differences between these two verbs through their semantic behaviors and complement types. Liu (2010) identified the semantic meanings and pattern differences of five synonymous - the adjectives chief, main, major, primary, and principal - using a corpus. Chief et al. (2000), on the other hand, used corpus data to examine the distributional patterns of the Chinese near-synonyms方便fang1bian4 and 便利bian4li4 (both mean “to be convenient”). Most studies on near-synonyms, regardless of the part-of-speech they analyzed, have so far arrived at the consensus that near-synonyms can be distinguished in terms of semantic and/or syntactic patterns. It has been stated that the lack of knowledge of the distinctions among near-synonyms could be one reason students use them incorrectly (Partington 1998, Hoey 2000, McEnery & Xiao 2006).

In the present paper, the English near-synonym pair listen (to) and hear was examined using different types of corpora.1 The overall purpose of this study was to first conduct a corpus-based analysis of listen (to) and hear and, later, to apply the information gathered from the corpora to a writing task. The research questions are as follows:
(a) How can a corpus-based study elicit similarities and differences between the verbs listen (to) and hear?
(b) How do the verbs listen (to) and hear behave linguistically when observed in a general corpus and in a writing task?

These research questions will be answered based on three sets of data:
  • WordNet 2.1, which contains the sense information of each word taken from two corpora (the Brown Corpus and Stephen Crane’s novella The Red Badge of Courage) (Landes et al. 1998);
  • a randomly selected sample from a native-speaker corpus (British National Corpus); and
  • a collection of data based on a writing task.
Among these data, the first two are data from native-speaker corpora.

As Ellis and Barkhuizen (2005) pointed out, one way to understand how language learners acquire a second language is to examine their language production. Our study examined the similarities and differences between the verbs listen (to) and hear produced in a guided writing task versus the corpus data. In addition, we also made use of WordNet (Miller et al. 1990, Fellbaum 1998) to examine the senses of the two verbs.2 The comparisons between the lexical resource and the corpora data should achieve the following results:
  • show how the uses of these two verbs may (or may not) differ in different corpora; and
  • provide an opportunity to compare the uses of these two verbs in a task based on which elicitations of these two verbs are intentional.
In the following section, the methodology employed will be discussed.


2 Methodology

Two main types of data were used, namely, native-speaker data and learner data. The native-speaker data were constituted by the analyses of senses in WordNet and the British National Corpus (BNC), while the learner data were comprised of students’ writings, which were collected through a designed activity.

Our methodology can be divided into three main steps:
  • The meanings of the verbs listen and hear as well as their respective sense frequencies were compared based on their senses provided by WordNet 2.1. This comparison of senses also ensured that these two verbs share at least one meaning so as to qualify as near-synonyms.
  • The analyses of the two verbs were then carried out based on 500 random examples taken from the BNC.3 The analyses were based on collocations, verb forms, and the meanings of the two verbs, and each instance was compared to the senses provided by WordNet 2.1. Statistical data on collocations and verb forms were extracted from Sketch Engine (Kilgarriff & Tugwell 2001), a lexical resource that enables collocational patterns to be displayed according to grammatical relations, such as subject, object, and modifier.
  • The data from the BNC were further compared to the elicited data collected from a writing task.

3 Analysis

3.1 Step One: WordNet

The search in WordNet 2.1 was conducted so as to investigate the senses of the verbs listen and hear and to further decide whether the two are near-synonyms. The senses and their respective sense frequencies are presented in Table 1 below. As mentioned above, the frequencies shown in Table 1 represent the frequencies provided by WordNet based on the occurrences of the two verbs in two corpora - the Brown Corpus and Stephen Crane’s novella The Red Badge of Courage. The numbers in Table 1 indicate how frequently these senses were found in these two corpora.

Listen
Hear
Senses
WordNet 2.1
Frequency
(in percent)
Senses
WordNet 2.1
Frequency
(in percent)
1
Hear with intention.
(e.g., Listen to the sound of this cello.)
60 (61.2%)
1
Perceive sound via the auditory sense.
275 (77.5%)
2
Listen and pay attention.
(e.g., Listen to your father.)
34 (34.7%)
2
Get to know or become aware of, usually accidentally.
(e.g., I heard that you have been promoted.)
60 (16.9%)
3
Pay close attention to.
(e.g., Listen to the advice of the old man.)
4 (4.1%)
3
Hear evidence by judicial process.
(e.g., The jury had heard all the evidence.)
12 (3.4%)



4
Receive a communication from someone.
(e.g., We heard nothing from our son for five years.)
8 (2.3%)



5
Listen and pay attention.
(e.g., We must hear the expert before we make a decision.)
N/A4
Total
98
(100.0%)
Total
355 (100.0%)

Table 1: The Distributions of Senses for the verbs Listen and Hear in WordNet 2.1

Table 1 shows that the verb listen (to) has three senses, in which the first sense - hear with intention - is most frequently used, constituting 61.2% of the total instances.5 The second sense of the verb listen - listen and pay attention (34.7%) - is also found among the senses of the verb hear (Sense 5), indicating that these two verbs have an overlapping meaning. The third sense pay close attention to is least frequently used - with only 4.1% of the total instances. On the other hand, the most frequent sense of the verb hear is perceive sound via the auditory sense with a coverage of 77.5% of the total instances displayed in WordNet. This is followed by the sense get to know or become aware of with 16.9% of the total instances. The other senses - with less than 5% each - are hear evidence by judicial process and receive a communication from someone. However, since the frequency of the sense listen and pay attention (Sense 5) is not available for the verb hear, comparisons between the two verbs cannot be made. This analysis shows that these two verbs are near-synonyms, with only one overlapping meaning. Most of the other meanings are different, which confirms the observations made by Chung and Ahrens (2008) and Taylor (2003). Cruse (1986: 267) also states that near-synonyms usually share some “central semantic traits” but differ in “peripheral traits".6 With the use of quantitative data, we showed what these traits are. The next step of our analysis was based on the BNC.

3.2 Step Two: The British National Corpus and the English Sketch Engine

The search using the English Sketch Engine resulted in 11,096 instances of listen appearing as verb forms (present tense, past tense, etc.).7 Hear, by contrast, has 34,609 instances as different verb forms. The corpus data from the BNC were analyzed in terms of collocations and verb forms, as reported in the following.

3.2.1 Collocations of Listen and Hear.

Based on corpus data from the BNC, it was found that listen almost never appeared in the transitive form as in *I listen words. In fact, both Dictionary.com and the Merriam-Webster online dictionary define the transitive meaning of listen as archaic:
Dictionary.com:verb (used with object). Archaic. to give ear to; hear. (http://www.dictionary.com/browse/listen?s=t;18-07- 2017.Merriam-Webster:archaic: to give ear to: hear. (https://www.merriam-webster. com/dictionary/listen; 18-07- 2017)
When searching for the prepositions of the verb listen in the BNC, we found the following results, which are shown in Table 2:
 Word
Observed Collocate Frequency
T-score Value
 to
6026
74.35
 for
242
10.11
 with
162
7.66
 without
18
3.22
 outside
10
2.79
 at
69
2.15
 through
14
1.86
Table 2: Top prepositions following listen in the British National Corpus

The collocation of listen with the top listed preposition to constituted most of the instances (6,026; 54.5%) in the 11,096 instances of listen as a verb.8 This means that most of the instances of the verb listen were intransitive. Comparatively, hear appeared in both transitive (I hear voices.) and intransitive forms (I heard from him lately.). Based on these findings, we compared the objects of listen (to) and hear (instead of listen and hear) in the analysis below. Table 3 shows the top 20 collocates under the grammatical relation of object for listen (to) and hear based on the results from Sketch Engine. Note that some noun phrases were not collected in Sketch Engine, as it provides collocates only at the word level. An example is given in (1) below:
(1) (...) but the number of people that don’t listen to what customers are saying to them
In Table 3 below, saliency is a measurement of significance in terms of the collocates used with a certain word in a certain grammatical relation.9 The overlapping objects (shaded in gray) in the top 20 collocates between listen (to) and hear are story, sound, voice, noise, and footstep, which appear in each list.10

Listen to
Hear
Collocates
Frequency
Saliency
Collocates
Frequency
Saliency
music
199
38.13
voice
718
43.38
radio
117
34.99
sound
497
43.31
tape
62
28.99
footstep
137
42.45
report
155
28.14
noise
229
38.4
conversation
54
26.64
scream
82
36.39
story
88
25.8
rumour
94
34.42
sound
55
22.66
news
253
32.53
voice
72
21.92
cry
94
31.62
lecture
24
20.42
shout
50
30.94
word
77
19.65
story
226
28.62
wind
33
19.62
click
35
28.42
noise
27
19.35
word
331
27.77
advice
37
18.74
murmur
30
27.28
song
18.46
bang
39
26.56
footstep
12
17.78
thud
23
26.28
commentary
12
17.65
bell
63
25.88
speech
29
17.06
whisper
33
25.77
recording
18
16.91
siren
26
24.93
breathing
10
16.24
knock
33
24.89
talk
26
15.8
rustle
17
24.86
Table 3: Top 20 Collocates for the object Relation of Listen to and Hear

As also found in the sense analysis in Table 1, perceive sound via the auditory sense was the most frequent sense of hear. Table 3 shows that voice, sound, footstep, and noise appear at the top of the list for hear, but they appear at different positions in the listen-to list. Similarly in Table 1, the most frequent sense of listen (to) was to hear with intention, and this is reflected in the top five objects - music, radio, tape, report, and conversation, which is information that listeners perceive with intention - in Table 1 .11 Furthermore, we also found that many of the objects of hear had a negative meaning (scream, rumour, cry, shout, and siren; in bold, and gunshots, explosion, roar and gunfire, which were not among the top 20 collocates), representing events that occur unexpectedly to the agent. When used with the verb listen to, the unintended meaning is lost, i.e., someone listens to gunshots, an explosion, a roar and gunfire on purpose, which is possible but not usual.

In fact, as near-synonyms, the verbs listen to and hear are sometimes used interchangeably in limited contexts where their meanings overlap, such as the substitutable meanings of the verbs listen to and hear in Table 3. In examples such as listen to the radio and hear on the radio, the noun radio denotes the event of broadcasting through the radio as a medium, not the physical radio itself. More examples of this are shown in (2) below. In these examples, the object (opinion) and the noun phrase (what people are saying) are underlined:
(2)  (a) I would love to listen to the opinion of other readers on this subject.      (b) I would love to hear the opinion of other readers on this subject.      (c) Before reaching your final decision, you must listen to what people are saying.      (d) Before reaching your final decision, you must hear what people are saying.
The examples in (2) show the same sense for two verbs, and this shared meaning is also the feature that defines these two verbs as a near-synonymous set. Unlike the sentences in (2) above, not all uses of listen to and hear in (3) below are interchangeable. Even though the same object (e.g., story) appears with both verbs, their meanings may not be identical because they can be interpreted differently, as shown in (3). In (3a), listen to stories means listen to narrated stories (with the intention of listening). By contrast, heard stories in example (3b) means receive messages (probably from unknown sources) unintentionally:
(3) (a) We used to love to listen to stories about the past of the family.
      (b) He had heard many stories about Yanto and knew he was a rough handful.
These two sentences clearly show a distinctive feature of these two verbs - intendedness in listen (to) is unmarked, while intendedness in hear is marked (Battistella 1990). The noun stories is therefore used in two different senses with listen to and hear, respectively, with (3a) referring to an account of incidents or events and (3b) referring to the metaphorical extension of stories to mean a widely circulated rumor (meanings taken from Merriam Webster Online, 18-07-2017; https://www.merriam-webster.com/dictionary/story).

In addition to the examples discussed above, listen to and hear may denote different senses in some other contexts where they become non-interchangeable. In other words, if the one verb is replaced by the other one, this may result in a change of meaning as shown in (4) below:
(4)  (a) You can listen to the radio while you’re working.      (b)?You can hear the radio while you’re working.      (c) He listened to the screams and bangs coming from Beatrice’s cottage.      (d) He heard screams and bangs coming from Beatrice’s cottage.
With regards to the examples (4a) and (4b), listen to the radio does not mean the same as hear the radio. These two uses differ in the intendedness of listen to, which is absent in hear (i.e., hear the radio means that someone hears some information by accident or without the intention of listening). Example (4c) is acceptable because the definite article (underlined in the examples) is added before screams and bangs. In this case, listened to has an undertone of "intendedness", meaning that the objects are definite as the listeners know what they are listening to. By contrast, the definite article may or may not be present in (4d) because hear does not have this restriction.

In (5) below, it is more natural to say (5a) than (5b) because the adverb attentively contradicts the unintended meaning of heard:
(5) (a) Helen listened attentively as Sophie revealed her new plan.      (b) *Helen heard attentively as Sophie revealed her new plan.
To further explain the similarities and differences of the two verbs as displayed in example (7), Table 4 provides the collocational data for the grammatical relation of modifier for the verbs listen (to) and hear:
Listen (to)
Hear
Collocates
Frequency
Saliency
Collocates
Frequency
Saliency
attentively
63
61.8
yesterday
289
51.1
intently
79
60.8
before
86
39.7
carefully
201
56.9
almost
79
27.9
sympathetically
12
29.3
distinctly
20
26.3
politely
15
28.7
clearly
53
25.9
hard
29
27.8
today
41
25.1
patiently
13
27.4
all
44
25
closely
20
22.8
once
46
24.6
please
22
22.6
aright
6
24.4
impassively
5
22
scarcely
18
23.4
Table 4: Top 10 Collocates for the modifier Relation of Listen (to) and Hear12

The data in Table 4 show that the verb listen (to) is followed by adverbials such as attentively, intently, carefully, sympathetically, impassively and closely, most of which show that the listener is in control and how the listening process is conducted, depending on the intention of the listener. Comparatively, the verb hear is usually followed by adverbials in terms of time (yesterday, before, today and once) and manner (distinctly, clearly, aright and scarcely).

Through Sketch Engine, terms coordinating with listen (to) and hear were also found. Listen (to) was found to often collocate with physical bodily actions such as sit, speak, talk, respond, stand and stay (in boldface), and only one perception verb (look) is shown in Table 5 (though the list is not exhaustive for listen (to)).13 On the contrary, hear only collocates with four verbs only, and two of them are perception verbs (see and smell) while the other two are mental verbs (determine and know):


Listen (to)

Collocates
Frequency
Saliency
Collocates
Frequency
Saliency
sit
49
29.4
stop
17
17.0
speak
29
24.4
look
29
16.8
talk
24
22.0
wait
12
15.7
learn
19
19.6
stand
12
13.3
respond
12
18.4
stay
8
12.6


 Hear

Collocates
Frequency
Saliency
Collocates
Frequency
Saliency
see
331
40.7
determine
21.0
19.8
smell
15
24.8
know
 9.0
 5.6
Table 5: Collocates for the relation of and/or for listen (to) and hear

Table 5 shows that listen (to) and hear collocate with certain verb types more often than others. The examples in (6) show some coordinating verbs with listen (to) and hear, respectively:
(6)  (a) Children learn to listen and speak through participation in projects.      (b) ?Children learn to hear and speak through participation in projects.      (c) As a very small child, I’d sit and listen while he read me the comic.      (d) ?As a very small child, I’d sit and hear while he read me the comic.      (e) ?Jason saw and listened nothing during the night.      (f) Jason saw and heard nothing during the night.      (g) ?Many elderly people can’t see or listen very well.      (h) Many elderly people can’t see or hear very well.
In (6), some of the examples are unnatural when listen (to) and hear are substituted for one another This, again, can be related to the unintendedness of hear because in examples such as (6b) and (6d), it is clear that when the agents speak or sit, their actions are conscious and intended. Therefore, the combination of hear and speak and sit and hear, in which an unintended action (hear) is combined with an intended action (speak/sit), is grammatical but semantically awkward. Similarly, when someone saw and heard nothing (6f), this means that the hearing process occurred without the intention of the agent. Nevertheless, the combination of saw and listened in (6e) violates this meaning because it hints at a purposive listening process which has no goal (nothing) and is therefore contradictory. A similar interpretation can be applied to (8g) because both see and hear can be unintentional, but the combination of see or listen represents a conflict of meaning. In the following section, the comparisons of verb forms for listen (to) and hear will be examined.

3.2.2 Verb Forms of Listen (to) and Hear.

Figure 1 shows how listen (to) and hear may also differ in their verb forms and how these verb forms may affect the meanings of these two verbs:

Figure 1: Distributions of Different Verbs Forms for Listen (to) and Hear14

Figure 1 shows that compared to hear, listen (to) appeared more frequently in the -ing form (29.3% vs. 6.4%) and in the finite base form (24.3% vs. 8.7%). Examples of these two verb forms are given in (7) below:
(7)  (a) The old man and the girl are listening attentively.      (b) I kept it up until I was certain you were not hearing a word.      (c) Listen to this, listen to this.      (d) Now hear this and hear it good.
The progressive form is usually employed for events which happen throughout a certain period of time; thus, the verb hear, which usually denotes an event that happens unexpectedly and unintendedly, is not commonly found in the -ing form.

In contrast, hear was encountered more frequently in the infinitive, the past tense and the past participle than listen, as is exemplified in (8):
(8) (a) Everyone always wanted to listen to what she had to say.      (b) We want to hear your views about any issue affecting the countryside.      (c) They stood in the darkness behind the door and listened to the foot steps.      (d) I heard someone coming towards the door.      (e) It is important that the advice of experts should be listened to.      (f) The voice of the novelist is heard continually in the speech of his characters.
The verb hear (29.1%) was slightly more frequently in the infinitive form than listen (24.6%). The verb hear in the past tense constituted 27.6% as compared to the verb listen in the past tense (15.2%). As a past participle, heard was far more frequent (26.9%) than listened (4.3%). In fact, we found that hear was usually followed by nominals (objects) which were negative in meaning (e.g. (9a) and (9c)). These nominals were often fronted (e.g. (9b) and (9d)) to emphasize what was being heard (underlined):
(9)  (a) He heard single gunshots from the house.
      (b) Occasional gunshots can still be heard.
      (c) It was like a nightmare. I heard a loud explosion just to my left.
      (d) explosions and mortar bombs were heard intermittently throughout the day.
This is one of the reasons why the verb hear is more frequently used in the passive form than the verb listen (to). Some examples of the verb listen (to) in the past participle form are given below:
(10) (a) It [music] can be played, listened to, read and written throughout the world…        (b) Children also need to be listened to and their point of view understood.
From the analyses of corpus data based on the British National Corpus through the Sketch Engine, at least six important differences between the verbs listen (to) and hear were identified.
  • The verb listen (to) was marked with intendedness while the verb hear was marked with unintendedness
  • The verb listen to was most frequently used to indicate that someone hears something with intention (e.g., listen to music / a report / the radio), and the verb hear was frequently used to refer to someone who perceives sound through the auditory sense (e.g., hear a voice / sound / noise).
  • Due to the undertone of unintendedness in the verb hear, it often collocated with words that had a negative connotation.
  • Since the objects of the verb hear were often negative, they were often fronted and as a result, more uses of past participle were found for he verb hear than for the verb listen (to). Comparatively, more progressive and finite base forms (especially in the imperative) were found for the verb listen (to) because both these forms allow purposive actions to take place with intention.
  • In terms of similarities, both verbs were used with objects which denote the production of physical sounds (e.g., music, song, footsteps, breathing, words, etc.).
  • In addition, both verbs were used with an event as in the phrases listen to the radio and hear on the radio, in which radio denotes an event rather than the device itself.

3.2.3 Step Three: Teaching the Verbs Listen to and Hear

3.2.3.1 Background

On the basis of our linguistic analysis of listen (to) and hear, we explained the two verbs to a group of students in an English class. Our purpose was to teach the specific meanings of the two verbs and explain to students how to distinguish them. The explanation took about forty minutes and students were tested in the following week. The following provides the background of the participants.

Thirty-nine undergraduate students from a national university in Taiwan participated in this task. Among these students, 12 were males, and 27 were females. Their average age was 20. The students were either freshmen or juniors, and they were paid for participating in this activity.
The activity consisted of a writing task and a subsequent vocabulary test. Students were asked to write a 100- to 150-word paragraph based on a series of four pictures adapted from the picture book Frog, Where Are You? (Mayer et al. 1969). This story is about a boy searching for his pet frog, in which the third picture (Figure 2) illustrates the boy putting his hand near his ear to search for the sound of the frog. Based on this illustration, students were expected to use either listen (to) or hear to describe this particular picture, and their language patterns were collected and examined.

Picture 1 (Mayer et al. 1969, 1) 


Picture 2 (Mayer et al. 1969: 3)




Picture 3 (Mayer et al. 1969: 23)    



Picture 4 (Mayer et al. 1969: 27)


Figure 2: Pictures Taken Frog, Where Are You? (Mayer et al. 1969)

The task began with a questionnaire about students’ personal information and language background. The main writing task, which included the instructions on the first page, the pictures on the second page, and a blank sheet of paper on the third page, was then distributed to students. The instructions were also explained by the instructor in Mandarin so as to ensure that students were clear about the task. Before the writing task began, students were allowed to ask questions for clarification regarding the task. After making sure that all the instructions were clear to students, the instructor then asked them to turn to the second page. Students were told to focus on the actions of the boy without being informed about the real purpose of the task. They then turned to the third page and started the writing task.

The whole process took about 50 minutes, and the writing task was completed within 30 minutes. After finishing the writing task, all the students received a vocabulary test, which was adapted from the vocabulary test by Redman (2002). There were ten multiple-choice questions on the test, and each question was accorded a score of ten points. Students were asked to hand in the test after five minutes. The vocabulary test was used as a reference in addition to collecting students’ self-reported English proficiency scores at the beginning of the activity (students had rated their own ability and provided their previous official examination scores or their English scores recognized by the university before their admission).

3.2.3.2 Results

The mean of students’ self-rated Mandarin proficiency was 6.8 on a scale ranging from 1 to 7, with 1 being the least proficient and 7 being the most proficient. The mean of students’ self-rated English proficiency was 4.6 The initial questionnaire also collected students’ self-reported scores on their English Basic Competency Test (BCT) for Junior High School Students and their Joint College Entrance Examination (JCEE). These scores were used as references to determine whether students should be dropped due to low English proficiency. The results showed that no student was dropped due to low English proficiency.

For the writing task, the instructor then transcribed all the writing samples into text files. The files were then run, using the AntConc Concordancer (Anthony 2005) to search for [listen*] and [hear*].15 Six out of the 39 students did not use the two verbs, mainly because they skipped the boys action in the third picture. They may have thought that this picture carried less important information than the other pictures. Ten students used both verbs listen (to) and hear, while the others used only one of the two verbs. Two examples are given in (11)16:
     (11) (a) Listen! I think I heard something. (S006)             (b) Therefore, he decided to search for it by listening to the frog’s sounds….     Suddenly, he heard some sounds from the timber on his left side. (S026)
In total, 15 instances (28.8%) of the verb listen (to) and 37 instances (71.2%) of the verb hear were found. Therefore given the picture with the boy putting his hand near his ear, students used more instances of hear than listen (to). Uses of both verbs are documented in (12):
(12) (a) The little boy puts his hand near his ear to hope to listen more clearly.          (S008)        (b) Tom was listening to the sounds the frog produced. (S026)        (c) They heard a voice of the frog. (S009)        (d) Suddenly, he heard some voices. (S015)
The examples in (12) match those from the above analysis, regarding the unintended meaning of the verb hear, which indicates that students were familiar with these two verbs.

The higher percentage of the verb hear was, however, predictable because, as shown in the previous analyses, the verb hear is more suitable for contexts expressing unintendedness, and the picture elicited a higher number of unintended uses. Among the students, 28.8% still preferred the intended use of listen (to) compared with hear, as in (12a) and (12b), both of which are purposive actions. Therefore, based on this activity, we found the uses of listen (to) and hear that were preferred by students when prompted by a picture. Most students chose to use hear to increase the suddenness of events in the story. For the verb listen (to), all of the instances (except one imperative) were intransitive. Hear, on the other hand, collocated with only three types of objects in the 37 instances, namely noise (3; 8.10%), sound (22; 59.46%), and voice (12; 32.43%). Among these, only eight instances (21.60%) were used with adjectives (slight, noisy, strange, delightful, and familiar). We had hoped that, at the undergraduate level, students would provide more sophisticated combinations when using these two verbs than they actually did.

To compare the results from the three resources, the distributions of the various meanings are presented in Table 6 below:  

WordNet
BNC
Listen (to) 
WordNet 2.1 Senses
Freq. (%)
Freq. (%)
Freq. (%)

1
Hear with intention. (e.g., Listen to the sound of this cello.)

60 (61.2%)
115 (23.0%)
14 (93.3%)

2
Listen and pay attention. (e.g., Listen to your father.)

34 (34.7%)
321 (64.2%)
1 (6.7%)

3
Pay close attention to. (e.g., Listen to the advice of the old man.)

4 (4.08%)
48 (9.6%)
N/A


Others
N/A
16 (3.2%)
N/A






Total
98 (100.0%)
500 (100.0%)
15 (100.0%)





WordNet
BNC
Student
Writing
Hear
WordNet 2.1 Senses
Freq. (%)
Freq. (%)
Freq. (%)

1
Perceive sound via the auditory sense.
275 (77.5%)
156 (31.2%)
37(100.0%)


2

Get to know or become aware of, usually accidentally. (e.g., I heard that you have been promoted.)

60 (16.9%)
278 (55.6%)
N/A

3
Hear evidence by judicial process. (e.g., The jury had heard all the evidence.)

12 (3.4%)
18 (3.6%)
 N/A

4
Receive a communication from someone. (e.g., We heard nothing from our son for five years.)

8 (2.3%)
14 (2.8%)
 N/A

5
Listen and pay attention. (e.g., We must hear the expert before we make a decision.)
N/A
28 (5.6%)
 N/A
Total
355 (100%)
500 (100%)
37 (100%)

Table 6: Distribution of senses in data from WordNet, BNC, and the writing task
 
To summarize the analysis of the writing task, we can say that most of the students were able to distinguish the near-synonym pair listen (to) and hear and use each of them in appropriate contexts when guided. They used the verb listen (to) so as to emphasize that something was paid close attention to and the verb hear to describe someone perceiving a sound, although the addition of adverbs may have made the unintentionality more subtle.


4 Discussion and Conclusion

In this paper, corpus data were used to examine the verbs listen (to) and hear in terms of their similarities and differences, as well as their semantic and syntactic distributional patterns. Similar to listen (to), the verb hear can also be used to describe that something is being paid attention to. However, such use was not frequently found in the native-speaker data, as native English speakers generally use the verb hear to refer to the act of perceiving a given sound via the auditory sense.

The linguistic analysis that we have carried out in this study has the potential to increase the awareness of English teachers, whose job generally comprises the explanation of different word meanings to students.

As Hunston & Feng (2002: 3) noted, teachers tend to teach language based on their own intuition, without providing a better explanation regarding the question of why a certain phrase is more appropriate in a particular context than in another one. Alternatively, by using a corpus, teachers can help learners gain access to a great amount of data, which can be processed and presented by “showing frequency, phraseology, and collocation” (Hunston & Feng 2002: 3). Our results can provide students with information about the distribution of lexical items (i.e., which senses of a word - like the verbs listen (to) and hear - are most frequently used or which verb forms are highest in percentage). This information will be valuable to language teachers because it provides the linguistic behaviors of the words compared.



Acknowledgements
This study was conducted under research grants from the Ministry of Science and Technology, Taiwan: 104-2420-H-004-034-MY2 and 106-2410-H-004-109-MY2. The author would like to thank the anonymous reviewers and the editor of JLLT for commenting on the previous version of this article. Tzu-Yun Tseng's help on the previous version was also appreciated.




References

Ahrens, K., H. Huang & Y. H. Chuang (2003). Sense and Meaning Facets in Verbal Semantics: A MARVS Perspective. In: Language and Linguistics 4 (2003), 468.

Anthony, L. (2005). AntConc: Design and Development of a Freeware Corpus Analysis Toolkit for the Technical Writing Classroom. In: Professional Communication Conference, 2005. IPCC 2005. Proceedings. International, IEEE, 729-737.

Battistella, E. L. (1990). Markedness: The Evaluative Superstructure of Language. Albany, NY: SUNY Press.

Baker, C. F., C. J. Fillmore & J. B. Lowe (1998). The Berkeley Framenet Project. In: Proceedings of the 17th International Conference on Computational Linguistics – Vol. 1. Association for Computational Linguistics, 86-90.

Biber, D., S. Conrad & R. Reppen (1998). Corpus Linguistics. Cambridge: Cambridge University Press.

Bolinger, D. L. M. (1977). Meaning and form (Vol. 1). London, UK: Longman.

Chief, L. C. (2000). What Can Near Synonyms Tell Us? In: Computational Linguistics and Chinese Language Processing 5 (2000) 1, 47-60.

Chung, S.-F. & K. Ahrens (2008). MARVS Revisited: Incorporating Sense Distribution and Mutual Information into Near-Synonym Analyses. In: Language and Linguistics: Lexicon, Grammar and Natural Language Processing 9 (2008) 2, 415-434.

Cruse, D. A. (1986). Lexical Semantics. Cambridge: Cambridge University Press.

Ellis, R. & G. P. Barkhuizen (2005). Analysing Learner Language. Oxford, UK: Oxford University Press.

Fellbaum, C. (Ed.) (1998). WordNet: An Electronic Lexical Database. Cambridge, MA: MIT Press.

Hoey, M. (2000). A World Beyond Collocation: New Perspectives on Vocabulary Teaching. In: Lewis, L. (Ed.) (2000). Teaching Collocations: Further Developments in the Lexical Approach. Boston: Heinle, 224-245.

Housen, A. (2002). A Corpus-based Study of the L2-acquisition of the English Verb System. In: Computer Learner Corpora, Second Language Acquisition and Foreign Language Teaching 6 (2002), 2002-2077.

Hunston, S. & Z. Feng (2002). Corpora in Applied Linguistics. Cambridge: Cambridge University Press.

Kennedy, G. (1998). An Introduction to Corpus Linguistics. London: Longman.

Kilgarriff, A. & D. Tugwell (2001). Word Sketch: Extraction and Display of Significant Collocations for Lexicography. Proceedings of the Workshop on Collocation: Computational Extraction, Analysis and Exploitation. 39th Annual Meeting of the Association for Computational Linguistics (ACL). Toulouse, France.

Landes, S., C. Leacock & R. I. Tengi (1998). Building Semantic Concordances. WordNet: An Electronic Lexical Database. Cambridge, MA: MIT Press, 199-216.

Liu, D. (2010). Is it a Chief, Main, Major, Primary, or Principal Concern? A Corpus-based Behavioral Profile Study of the Near-synonyms. In: International Journal of Corpus Linguistics 15 (2010) 1, 56-87.

Lyons, J. (1968). Introduction to Theoretical Linguistics. London: Cambridge University Press.

Mayer, M. et al. (1969). Frog, Where Are You?. New York: Dial Press.

McEnery, A. & Z. Xiao (2006). Collocation, Semantic Prosody and Near Synonymy: A Cross-linguistic Perspective. In: Applied Linguistics 27 (2006) 1, 103-199.

Miller, G. A. et al. (1990). Introduction to Wordnet: An On-line Lexical Database*. In: International Journal of Lexicography 3 (1990) 4, 235-244.

Partington, A. (1998). Patterns and Meanings: Using Corpora for English Language Research and Teaching. Philadelphia, PA: John.

Redman, S. (2002). English Vocabulary in Use: Pre-intermediate and Intermediate. London: Cambridge University Press.

Taylor, J. R. (2003). Near Synonyms as Co-extensive Categories: ‘High’ and ‘Tall’ Revisited. In: Language Sciences 25 (2003) 3, 263-284.



Author:
Siaw-Fong Chung, Ph.D.
Associate Professor 
Department of English
National Chengchi University
No. 64, Sec. 2, ZhiNan Road
Taipei City 11605, Taiwan


1 The preposition to is sometimes needed to establish listen as a near-synonym with hear. As such, we can already predict that these two verbs are not perfect synonyms because they differ syntactically. A more in-depth analysis will be provided in later sections of this paper.
2 WordNet is a lexical database that provides the semantic relations (such as synonyms, antonyms, hypernyms, and hyponyms) of lexical items. It is available at http://wordnet.princeton.edu/. The term sense refers to the meaning entry listed in WordNet. Since our analysis of the meanings of listen (to) and hear is based on their senses, the terms sense and meaning are used interchangeably in this article.
3 The British National Corpus is a 100-million-word collection of written and spoken texts. For more information, cf. http://www.natcorp.ox.ac.uk/.
4 When “N/A” is shown, it is either because this sense never occurred in the corpora or because this sense was added after the sense analysis had been carried out.
5 The total numbers in Table 1 were computed by the authors, who assumed that the respective frequency for each sense also reflected the total number of tokens found for listen and hear in the two corpora examined.
6 However, as shown in Table 1, the “central” trait may not be the most frequent trait.
7 Listen and listen to are both near-synonyms of the verb hear, but in some corpora, the search for listen to is not allowed (such as the search for collocations in Sketch Engine).
8 An asterisk (*) serves as a wildcard to extract any verb forms beginning with listen (to) and hear (including -ing, -ed, etc.). However, when an asterisk is used in an example sentence, such as *I listen music, this means that the sentence is grammatically incorrect. A question mark appearing before a sentence means that the sentence is unnatural but not ungrammatical.
9 However, saliencies are list-independent and, thus, cannot be compared across lists (such as between listen (to) and hear), but they do indicate the rank of importance among the collocates in the same list.
10 The total number of collocates for the objects of hear does not amount to 34,609, which is the total number of instances for hear in the whole British National Corpus, because some instances of hear were intransitive.
11 Therefore, it can also be postulated that sense frequency can be obtained by observing the ranking of the collocates, which is arranged according to saliency. Predicting sense frequency is an interesting area of research that can be carried out through collocational analyses.
12 The collocates in Tables 3 and 4 can apply to both listen and listen to. Thus, the notation listen (to) is used in the titles and respective columns of both tables.
13  Both respond and stay can also be mental verbs.
14 The results in Figure 1 only provide the forms of the verbs (e.g., listening, listened). We still do not know how many of these verb forms are used with to (e.g., listening to vs. listening; listened to vs. listened).
15 See footnote 8 for the use of asterisks. As the asterisk is a wildcard, the search might return irrelevant results such as the form heart. Irrelevant results were removed manually. Noun forms such as listener were also excluded in our results because only the verbs were included for further analysis.

16 Each text was assigned a number (e.g., “S006”) as all texts were saved anonymously.
As the texts were written by students, they may contain some language errors.