YU Haiyan, HU Qinyou

(Merchant Marine College, Shanghai Maritime University, Shanghai 201306, China)

Analysis on semantic relations of special structures of noun phrases in marine engineering English

YU Haiyan, HU Qinyou

(Merchant Marine College, Shanghai Maritime University, Shanghai 201306, China)

In order to analyze the semantic structures and semantic relations between the noun phrases in the Marine Engineering English (MEE) so as to make marine engineers better understand specialized vocabulary, SPSS is applied to compare the self-built MEE corpus and the online FROWN corpus by Chi-square test. Based on the theory of 16 classifications of semantic relations of noun + noun (see BIBER D,etal.LongmangrammarofspokenandwrittenEnglish[M]. Beijing: Foreign Language Teaching and Research Press, 2000: 589-596), the semantic structures and semantic relations are analyzed for the self-built MEE corpus. The conclusions are as follows. (1)There is a significant difference between the MEE and common English in noun phrases. (2)The phrases of “noun + noun” structure in the MEE covers all 16 classifications of semantic relations, where the semantic relations of more than 10% is for the purpose of Purpose, Identity and Specialization; the semantic relations of “noun + noun + noun” structure phrases is grouped into 3 types of tree structures, and for each type, the distribution frequency of the semantic role is different.

marine engineering English; noun phrase; semantic relation; corpus; SPSS

0 Foreward

Maritime English - the entirety of all those means of the English language being used as a device for communication within the maritime community-internationally contributes to the safety of shipping and organization of the seaborne business.[1]It is of great necessity to learn maritime English for any member who benefits from or associated with the shipping industry. Especially, for the Marine Engineering English (MEE), it is an indispensable part of maritime English.

The MEE belongs to the ESP (English for Specific Purposes) used by marine engineers. Normally, the features of the modern English have been studied in detail from all respects. In this paper, the features of the MEE are analyzed through taking noun phrases as examples especially. ZHANG[2]signified that the noun phrases in the MEE are always constructed with tediously long elements, which hampers the beginners to acquire them. QUIRKetal[3]simply classified the noun phrases into two types: basic noun phrases and complex noun phrases. One type of the complex noun phrases - noun + head, for example, (some expensive) hotel furniture - has two nouns in it. THOMPSON[4]explained the nominal groups’ function in the language. More specifically, RADFORD[5]used the tree structure to demonstrate the internal structure of noun phrases. BIBERetal[6]gave 16 classifications to tell the mutual semantic relations in noun+noun sequences. Prior to telling the semantic relations, these roles of any nominal word are called thematic roles. SAEED[7]generalized these thematic roles into 9 types. XUE[8]demonstrated that the semantic relations or meanings of noun phrases cannot be clearly and correctly understood without precise analysis of their constituent structures. GIRJUetal[9]labeled complex nominals and noun phrases by means of semantics scattering automatically and found out a list of 35 semantic relations to classify noun phrases.

As a new research sector, the study of noun phrases in the MEE is conducted by more scientific and precise approaches with collocation rules to define their mutual semantic relations. In terms of the investigation of the distribution of the different noun phrases with various numbers of nouns in them (the structures of “noun+noun”, “noun+noun+ noun”,…) in the modern English corpus and the MEE corpus, the following issues are discussed and tackled in this paper: (1) Are there some differences in the noun phrases between the modern English and the MEE? (2) How are the different noun phrases with various numbers (2, 3) of noun in them distributed in the modern English and the MEE? (3) For the MEE, what are the specific semantic relations between and among the nouns in the structures of “noun+noun”, “noun+noun+noun”, …

1 Research method

1.1 Theory of corpus-based linguistics and statistics

BIBERetal[10]stated that the text corpora can obtain large databases of naturally-occurring discourse, which enables analysis of a scope not otherwise feasible. ABNEY[11]illustrated that in 1996 no one can profess to be a computational linguist without a passing knowledge of statistical methods, like SPSS. HUNSTON[12]highlighted the function of corpora that via collecting and storing data they can present concordance lines and calculate frequencies and give quantitative support to the study.

1.2 Corpora used in the study: FROWN online corpus and self-built MEE corpus

BIBERetal[13]specified that the features vary in different registers, for example, in speech and written English applicable for ESP education. Two corpora are involved in this study: the FROWN corpus as a controlled corpus is a representative of modern English while the newly-built MEE corpus functions as the study corpus.

The FROWN corpus contains 500 texts of around 2 000 words each, which is distributed across 15 text categories. The MEE corpus contains more than 200 000 words with all materials including introduction manuals of sorts of machinery. Each word in the two corpora is labeled with the software TreeTagger, and sequences ended with -NN/NNS/NP/NPS as symbols of noun phrases are selected.

1.3 Data analysis

To test whether there are significant differences between the FROWN and the MEE corpora, Chi-square test can be used to compare the frequencies with two categories or more.[14]Table 1 contains the data from two corpora to be tested.

Table 1 Frequencies of four categories of noun phrases

After inputting the data above, the Chi-square test can also output the results in terms of SPSS. In terms of the formuladf=n-1=1(n=2), the calculated value 5.187 0 is much greater than the critical one 3.841 5. Therefore, there is a significant difference between the two categories from two corpora at the 5% level.

2 Results and discussions

2.1 Distribution and semantic relations in “noun+noun” structure in MEE corpus

It is known that there are 16 classifications of semantic relations in the “noun+noun” structure.[6]After the analysis on nearly 500 examples in the MEE corpus, the distribution of these 16 classifications can be clearly realized. Therefore, Fig.1 is produced without containing the error data (13 proper nouns and 7 wrongly labeled words).

Fig.1 indicates that the data from the MEE corpus covers the noun+noun classifications as many as the modern English (totally 16 classifications). However, in the MEE corpus, there are 3 classifications (classification 2, 3 and 15) with proportions higher than 10%.

Classification 2 covers 27.40% with the highest frequency. Classification 2 means that the two nouns have a mutual relation called purpose. In this type of “noun+noun” structure, the former noun is for the purpose of the latter one, and the latter noun is used for the former one.[6]“Storage tank” belongs to this type. Here, “tank” in the MEE has the special meaning: a large (usually metallic) vessel for holding gases or liquids. Therefore, in the noun phrase “storage tank”, “tank” has a purpose to store gases or some liquids.

Fig.1 Sixteen classifications and their proportions in MEE corpus

Classification 15 (the specialization relation between two nouns) covers 22.90% with the second highest frequency. “Water temperature” demonstrates that the “temperature” is restricted to water (not other substances). Hence, in this phrase, the “temperature” is specialized for water.

Classification 3 (the identity relation between two nouns) covers 12.20% with the third highest frequency. This classification intends to show that the noun can be distinguished from others through its identity. For example, “suction stroke” is distinguished from other strokes, like the compression stroke, expansion stroke, etc. Hence, the “suction” has become the identity of “stroke”.

2.2 Distribution and semantic relations in “noun+noun+noun” structure in MEE corpus

After collecting and analyzing the data of “noun+noun+noun” structure in the MEE corpus, 3 basic constitutional analysis tree structures can be produced. Table 2 shows their proportions in this corpus.

According to the calculated output, it is evident that tree 1 accounts for 59.3% of the noun phrases with a “noun+noun+noun” structure, but there are still 5 different sub-branches. The initial two nouns together play a semantic role to modify the last noun as follows: agent, theme, instrument, complement and partitive.

Table 2 Three tree structures of constitutional analysis and their proportions in MEE corpus

Tree 2 accounts for 36.6% in the MEE corpus with a “noun+noun+noun” structure. Some semantic roles still take part in this tree structure, like source, location, beneficiary and patient. Tree 3 accounts for the least proportion of 4.1%. This classification has a head, but the other two elements stand independently.

2.3 Summary

As the outcome of this study, there exist obvious differences in the noun phrases between the FROWN corpus and the MEE corpus. The features of the special structures with “noun+noun” and “noun+noun+noun” in MEE corpus are found as follows.

Firstly, with Chi-square test, there is a significant difference among the 4 categories of the noun phrases in the FROWN corpus and the MEE corpus as well as for the structures of “noun+noun” and “noun+noun+noun”. Secondly, in the MEE corpus, the structure of “noun+noun” covers all the 16 classifications of semantic relations, but only 3 of them have a higer frequency(>10%), which are classification 2 (purpose, 27.40%), classification 15 (specialization, 22.90%) and classification 3 (identity, 12.20%), respectively. Thirdly, for the special structure of “noun+noun+noun” in the MEE corpus, there are 3 classifications of tree structures: tree 1 (59.3%) constructs in the way that the first two nouns have a closer semantic relation; in tree 2 (36.6%), the first noun acts as the head, and the following two nouns together have a closer semantic relation to modify the head; in tree 3 (4.1%), both the first and second nouns have an independent function to modify the head. Fourthly, in tree 1, 5 different sub-branches exist from a viewpoint of the detailed semantic function and semantic roles as agent, theme, instrument, complement and partitive.In tree 2, some other semantic roles still take part in this tree structure, like source, location, beneficiary and patient. In tree 3, each element is independent relatively. Therefore, their semantic functions or roles are flexible but not limited to certain types.

3 Conclusions

According to this study, the types of the noun phrases of the Marine Engineering English (MEE) are capable of being identified under specific rules, which will facilitate the maritime English instructors to clarify the internal semantic meanings and relations so as to familiarize the students with the terminology noun phrases in the MEE. With insufficient knowledge of the internal semantic meanings of these terminology noun phrases, the marine engineering students find it too difficult to understand and memorize these words. Consequently, these kinds of inadequate knowledge will attribute to some horrible accidents. Therefore, in the process of the teaching, the types of the noun phrases with a higher frequency can be highlighted for the purpose of effective learning.

As the scope of the database is confined to 20 000 words, only the structures with two and three nouns are investigated. Consequently, the study of all special structures of noun phrases has not been conducted in detail. Based on an extended database, the further study on the structures with four, five or more nouns can also be investigated.

[1]TRENKNER P. Maritime English - an attempt of an imperfect definition[C]. Proceedings of 2nd IMLA Workshop on Maritime English in Asia (WOME 2A). Dalian, 2000: 1-8.

[2]張曉峰. 輪機(jī)英語(yǔ)名詞性詞組語(yǔ)義和搭配分析[J]. 大連海事大學(xué)學(xué)報(bào)(社會(huì)科學(xué)版), 2010, 9(3): 106-109.

[3]QUIRK R, GREENBAUM S, LEECH G,etal. A comprehensive grammar of the English language[M]. Cambridge: Cambridge University Press, 1987, 9: 109-111.

[4]THOMPSON G. An introduction to functional grammar[M]. Beijing: Foreign Language Teaching and Research Press, 2000: 179-185.

[5]RADFORD A. Transformational grammar: a first course[M]. Beijing: Foreign Language Teaching and Research Press, 2000: 372-238

[6]BIBER D, JOHANSSON S, LEECH G,etal. Longman grammar of spoken and written English[M]. Beijing: Foreign Language Teaching and Research Press, 2000: 589-596.

[7]SAEED J I. Semantics[M]. Beijing: Foreign Language Teaching and Research Press, 2000: 140-141.

[8]薛慕煊. 英語(yǔ)名詞短語(yǔ)中的若干語(yǔ)義問題[J]. 上海大學(xué)學(xué)報(bào)(社會(huì)科學(xué)版), 1990(1): 74-80.

[9]GIRJU R, MOLDOVAN D, BADULESCU A,etal.Models for the semantic classification of noun phrases[J]. Stroudsburg: Association for Computational Linguistics, 2004:60-67.

[10]BIBER D, CORAD S, REPPEN R. Corpus-based approaches to issues in applied linguistics[M]. Oxford: Oxford University Press, 1994, 15(2):169-189.

[11]ABNEY S. Statistical methods and linguistics[D]. Cambridge: University of Tubingen, 1996:1-24.

[12]HUNSTON S. Corpora in applied linguistics[M]. Xi’an: World Publishing Corporation, 2006: 3-13.

[13]BIBER D, CORAD S, REPPEN R. Corpus linguistics[M]. Beijing: Foreign Language Teaching and Research Press, 2000: 135-139.

[14]周世界. 語(yǔ)言統(tǒng)計(jì)學(xué)[M]. 大連: 大連海事大學(xué)出版社, 2004: 116-118.

(Editor JIA Qunping)

輪機(jī)英語(yǔ)中名詞短語(yǔ)特殊結(jié)構(gòu)的語(yǔ)義關(guān)系分析

于海燕，胡勤友

(上海海事大學(xué) 商船學(xué)院， 上海 201306)

為分析輪機(jī)英語(yǔ)中名詞短語(yǔ)之間的語(yǔ)義結(jié)構(gòu)和語(yǔ)義關(guān)系，以便輪機(jī)員更好地理解專業(yè)詞匯，運(yùn)用SPSS軟件，通過卡方檢測(cè)比較自建MEE語(yǔ)料庫(kù)和線上FROWN語(yǔ)料庫(kù).根據(jù)名詞+名詞16種語(yǔ)義關(guān)系理論(參見BIBER D,etal.LongmangrammarofspokenandwrittenEnglish[M]. Beijing: Foreign Language Teaching and Research Press, 2000: 589-596)，分析自建MEE語(yǔ)料庫(kù)中名詞短語(yǔ)之間的語(yǔ)義結(jié)構(gòu)和語(yǔ)義關(guān)系.得出結(jié)論：(1)輪機(jī)英語(yǔ)和通識(shí)英語(yǔ)中的名詞短語(yǔ)存在顯著性差異.(2)輪機(jī)英語(yǔ)中“名詞+名詞”結(jié)構(gòu)短語(yǔ)覆蓋了全部16種語(yǔ)義關(guān)系，其中超過10%的為目的關(guān)系、識(shí)別關(guān)系和專有屬性關(guān)系；“名詞+名詞+名詞”結(jié)構(gòu)短語(yǔ)的語(yǔ)義關(guān)系可分為3種類型的樹結(jié)構(gòu)，每種類型的語(yǔ)義角色的分布頻率不同.

輪機(jī)英語(yǔ)； 名詞短語(yǔ)； 語(yǔ)義關(guān)系； 語(yǔ)料庫(kù)； SPSS

10.13340/j.jsmu.2017.00.018

1672-9498(2017)01-0090-05

date：2016-01-21 Revised date：2016-10-13

Cultivation Plan of Shanghai Higher Education Teachers

H313; U664 Document code： A

Biography：YU Haiyan (1985-), female, Baotou of Inner Mongolia, lecturer， research area is maritime English, (E-mail)yuhy@shmtu.edu.cn; HU Qinyou (1974-), male, Shucheng of Anhui province, professor, PhD, research area is shipping and maritime information processing, (E-mail)qyhu@shmtu.edu.cn