DOCSLIB.ORG

Identifying False Friends Between Closely Related Languages

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Identifying False Friends Between Closely Related Languages Identifying False Friends between Closely Related Languages Nikola Ljubesiˇ c´ Darja Fiserˇ Faculty of Humanities and Social Sciences Faculty of Arts University of Zagreb University of Ljubljana Ivana Luciˇ ca´ 3 Askerˇ cevaˇ 2 10000 Zagreb, Croatia 1000 Ljubljana, Slovenija [email protected] [email protected] Abstract and have been almost exclusively limited to par- In this paper we present a corpus-based ap- allel data (Inkpen et al., 2005; Nakov and Nakov, proach to automatic identification of false 2009). In this paper we tackle the problem of auto- friends for Slovene and Croatian, a pair matically identifying false friends in weakly com- of closely related languages. By taking parable corpora by taking into account the distri- advantage of the lexical overlap between butional and frequency information collected from the two languages, we focus on measuring non-parallel texts. the difference in meaning between iden- tically spelled words by using frequency Identifying false friends automatically has the and distributional information. We ana- same prerequisite as the problem of detecting lyze the impact of corpora of different ori- cognates – identifying similarly (and identically) gin and size together with different associ- spelled words between two languages, which is far ation and similarity measures and compare from trivial if one takes into account the specificity them to a simple frequency-based base- of inter-language variation of a specific language line. With the best performing setting pair. In this contribution we focus on the prob- we obtain very good average precision of lem of false friends on two quite similar languages 0.973 and 0.883 on different gold stan- with a high lexical overlap – Croatian and Slovene dards. The presented approach works on – which enables us to circumvent the problem of non-parallel datasets, is knowledge-lean identifying similarly spelled words and use identi- and language-independent, which makes it cal words only as the word pair candidate list for attractive for natural language processing false friends. tasks that often lack the lexical resources and cannot afford to build them by hand. Our approach to identifying false friends relies on two types of information extracted from cor- 1 Introduction pora. The first one is the frequency of a false friend False friends are words in two or more languages candidate pair in the corresponding corpora where that are orthographically or semantically similar the greater the difference in frequency, the more but do not have the same meaning, such as the certain one can be that the words are used in dif- noun burro, which means butter in Italian but don- ferent meanings. The second information source is key in Spanish (Allan, 2009). For that reason, they the context from corresponding corpora where the represent a dangerous pitfall for translators, lan- context dissimilarity of the two words in question guage students as well as bilingual computer tools, is calculated through a vector space model. such as machine translation systems, which would all benefit greatly from a comprehensive collection The paper is structured as follows: in Section 2 of false friends for a given language pair. we give an overview of the related work. In Sec- False friends between related languages, such tion 3 we describe the resources we use and in Sec- as English and French, have been discussed by tion 4 we present the gold standards used for eval- lexicographers, translators and language teachers uation. Section 5 describes the experimental setup for decades (Chacon´ Beltran,´ 2006; Granger and and Section 6 reports on the results. We conclude Swallow, 1988; Holmes and Ramos, 1993). How- the paper with final remarks and ideas for future ever, they have so far played a minor role in NLP work. 69 Proceedings of the 4th Biennial International Workshop on Balto-Slavic Natural Language Processing, pages 69–77, Sofia, Bulgaria, 8-9 August 2013. c 2010 Association for Computational Linguistics 2 Related Work most similar but contextually most dissimilar word pairs. Automatic detection of false friends was initially The feature weighting used throughout the re- limited to parallel corpora but has been extended lated work is mostly plain frequency with one to comparable corpora and web snippets (Nakov et case of using TF-IDF (Nakov and Nakov, 2007) al., 2007). The approaches to automatically iden- whereas cosine is the most widely used similar- tify false friends fall into two categories: those that ity measure (Nakov and Nakov, 2007; Nakov and only look at orthographic features of the source Nakov, 2009; Schulz et al., 2004) while Mitkov and the target word, and those that combine ortho- et al. (2007) use skew divergence which is very graphic features with the semantic ones. similar to Jensen-Shannon divergence. Orthographic approaches typically rely on com- The main differences between the work we re- binations of a number of orthographic similarity port on in this paper and the related work are: measures and machine learning techniques to clas- sify source and target word pairs to cognates, false 1. we identify false friends on a language pair friends or unrelated words and evaluate the differ- with a large lexical overlap – hence we can ent combinations against a manually compiled list look for false friends only among identically of legitimate and illegitimate cognates. This has spelled words, such as boja, which means been attempted for English and French (Inkpen et buoy in Slovene but colour in Croatian, and al., 2005; Frunza and Inkpen, 2007) as well as not among similarly spelled words, such as for Spanish and Portuguese (Torres and Alu´ısio, the Slovene adjective bucenˇ (made of pump- 2011). kins and noisy) and its Croatian counterpart Most of the approaches that combine ortho- bucanˇ (only noisy); graphic features with the semantic ones have been 2. we inspect multiple association and similarity performed on parallel corpora where word fre- measure combinations on two different cor- quency information and alignments at paragraph, pora pairs, which enables us to assess the sta- sentence as well as word level play a crucial role at bility of those parameters in the task at hand; singling out false friends, which has been tested on Bulgarian and Russian (Nakov and Nakov, 2009). 3. we work on two different corpora pairs which Work on non-parallel data, on the other hand, of- we have full control over (that is not the case ten treats false friend candidates as search queries, with web snippets), and are therefore able to and considers the retrieved web snippets for these examine the impact of corpus type and corpus queries as contexts that are used to establish the size on the task; degree of semantic similarity of the given word 4. we use three categories for the identically pair (Nakov and Nakov, 2007). spelled words: Apart from the web snippets, comparable cor- pora have also been used to extract and clas- (a) we use the term true equivalents (TE) sify pairs of cognates and false friends between to refer to the pairs that have the same English and German, English and Spanish, and meaning and usage in both languages French and Spanish (Mitkov et al., 2007). In (e.g. adjective bivsiˇ , which means for- their work, the traditional distributional approach mer in both languages), is compared with the approach of calculating n- (b) the term partial false friends (PFF) de- nearest neighbors for each false friend candidate in scribes pairs that are polysemous and the source language, translating the nearest neigh- are equivalent in some of the senses but bors via a seed lexicon and calculating the set in- false friends in others (e.g. verb drazitiˇ , tersection to the N nearest neighbors of the false which can mean either irritate or make friend candidate from the target language. more expensive in Slovene but only irri- A slightly different setting has been investigated tate in Croatian), and by Schultz et al. (2004) who built a medical do- (c) we use the term false friends (FF) for main lexicon from a closely related language pair word pairs which represent different (Spanish-Portuguese) and used the standard distri- concepts in the two languages (e.g. noun butional approach to filter out false friends from slovo, which means farewell in Slovene cognate candidates by catching orthographically and letter of the alphabet in Croatian) 70 By avoiding the problem of identifying relevant 4 Gold Standards similarly spelled words prior to the identification The gold standards for this research were built of false friends, in this paper we focus only on the from identically spelled nouns, adjectives and latter and avoid adding noise from the preceding verbs that appeared with a frequency equal or task. higher than 50 in the web corpora for both lan- 3 Resources Used guages. The false friend candidates were categorized in In this paper we use two types of corpora: the three categories defined in Section 2: false Wikipedia corpora (hereafter WIKI) which have friends, partial false friends and true equivalents. gained in popularity lately because of their sim- Manual classification was performed by three ple construction and decent size and web corpora annotators, all of them linguists. Since identify- (hereafter WAC) which are becoming the standard ing false friends is hard even for a well-trained lin- for building big corpora. guist, all of them consulted monolingual dictionar- We prepared the WIKI corpora from the dumps ies and corpora for both languages before making of the Croatian and Slovene Wikipedias by ex- the final decision. tracting their content, tokenizing and annotat- The first annotation session was performed by a ing them with morphosyntactic descriptions and single annotator only. Out of 8491 candidates, he lemma information.
Load more

Recommended publications
  • Vlatko Broz Faculty of Humanities and Social Sciences, Zagreb [email protected]
    UDK 81’373.42 81’373.612.2 81’23:165.194 Izvorni znanstveni ~lanak Prihva}eno za tisak 14. 10. 2008. Vlatko Broz Faculty of Humanities and Social Sciences, Zagreb [email protected] 'LDFKURQLF ,QYHVWLJDWLRQV RI )DOVH )ULHQGV A number of internationalisms originating from Latin have different meanings in English from their cognates in other European languages. This paper presents the results of an in- vestigation of words such as actually, eventually, etiquette, fabric, billion, chef, preservative, sensible, sympathetic, biscuit in Oxford English Dictionary. All these words have retained more or less the same meaning in German, Spanish and Croatian (each taken in this re- search as a representative of three major linguistic branches of the Indo–European language family), whereas English shows a departure from the original meaning. The causes for the shift in meaning are explained through the cognitive mechanisms of metonymy and meta- phor and a tentative period is determined when these semantic changes were under way in the course of the development of the English language. ,QWURGXFWLRQ Although the topic of false friends has been exploited for almost a century (Koessler and Derocquigny 1928, Ivir 1968, Nilsen 1977, Chamizo Domínguez and Nerlich 2002)2, ever since Koessler and Derocquigny gave them this name back in 1928, there is still new light that can be thrown on this topic, espe- 1 This is a revised version of a paper read in the general section of the 10th International Cognitive Linguistics Conference 2007 (July 15–21) in Krakow, Poland. 2 Exploring the topic of false friends also has a long tradition in Croatia.
    [Show full text]
  • ED311449.Pdf
    DOCUMENT RESUME ED 311 449 CS 212 093 AUTHOR Baron, Dennis TITLE Declining Grammar--and Other Essays on the English Vocabulary. INSTITUTION National Council of Teachers of English, Urbana, Ill. REPORT NO ISBN-0-8141-1073-8 PUB DATE 89 NOTE :)31p. AVAILABLE FROM National Council of Teachers of English, 1111 Kenyon Rd., Urbana, IL 61801 (Stock No. 10738-3020; $9.95 member, $12.95 nonmember). PUB TYPE Books (010) -- Viewpoints (120) EDRS PRICE MF01/PC10 Plus Postage. DESCRIPTORS *English; Gr&mmar; Higher Education; *Language Attitudes; *Language Usage; *Lexicology; Linguistics; *Semantics; *Vocabulary IDENTIFIERS Words ABSTRACT This book contains 25 essays about English words, and how they are defined, valued, and discussed. The book is divided into four sections. The first section, "Language Lore," examines some of the myths and misconceptions that affect attitudes toward language--and towards English in particular. The second section, "Language Usage," examines some specific questions of meaning and usage. Section 3, "Language Trends," examines some controversial r trends in English vocabulary, and some developments too new to have received comment before. The fourth section, "Language Politics," treats several aspects of linguistic politics, from special attempts to deal with the ethnic, religious, or sex-specific elements of vocabulary to the broader issues of language both as a reflection of the public consciousness and the U.S. Constitution and as a refuge for the most private forms of expression. (MS) *********************************************************************** Reproductions supplied by EDRS are the best that can be made from the original document. *********************************************************************** "PERMISSION TO REPRODUCE THIS MATERIAL HAS BEEN GRANTED BY J. Maxwell TO THE EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC)." U S.
    [Show full text]
  • Recognizing Cognates and Interlingual Homographs: Effects of Code Similarity in Language-Specific and Generalized Lexical Decision
    Memory & Cognition 2004, 32 (4), 533-550 Recognizing cognates and interlingual homographs: Effects of code similarity in language-specific and generalized lexical decision KRISTIN LEMHÖFER and TON DIJKSTRA University of Nijmegen, Nijmegen, The Netherlands In four experiments, we investigated how cross-linguistic overlap in semantics, orthography, and phonology affects bilingual word recognition in different variants of the lexical decision task. Dutch– English bilinguals performed a language-specific or a generalized lexical decision task including words that are spelled and/or pronounced the same in English and in Dutch and that matched one-language con- trol words from both languages. In Experiments 1 and 3, “false friends” with different meanings in the two languages (e.g., spot) were presented, whereas in Experiments 2 and 4 cognates with the same meanings across languages (e.g., film) were presented. The language-specific Experiments 1 and 2 replicated and qualified an earlier study (Dijkstra, Grainger, & Van Heuven, 1999). In the generalized Experiment 3, par- ticipants reacted equally quickly on Dutch–English homographs and Dutch control words, indicating that their response was based primarily on the fastest available orthographic code (i.e., Dutch). In Experi- ment 4, cognates were recognized faster than English and Dutch controls, suggesting coactivation of the cognates’ semantics. The nonword results indicate that the bilingual rejection procedure can, to some ex- tent, be language specific. All results are discussed within the BIAϩ (bilingual interactive activation) model for bilingual word recognition. In a foreign country where an unfamiliar language is nition. Through such words, a wealth of studies in the last spoken, words such as hotel, taxi, and café can often still decade have revealed that during the initial stages of word be recognized because they possess the same or a similar identification by bilinguals, word candidates from several spelling and meaning across languages.
    [Show full text]
  • Phraseological False Friends in English and Slovene and the Metaphors Behind Them Marjeta Vrbinc University of Ljubljana, Slovenia
    Phraseological False Friends in English and Slovene and the Metaphors behind them Marjeta Vrbinc University of Ljubljana, Slovenia The interest in false lexical equivalence reflects the interest in language contact, the observation of which always leads to the conclusion that formally identical and similar words and word combinations in different languages do not necessarily overlap semantically. Dictionaries of false friends deal with one- word lexical items, but false-friend relationship can also be established in phraseology. The aim of this paper is to look at phraseological components of English and Slovene lexicons with a view to identifying and describing the false semantic equivalence between idioms in these two languages. When studying false lexical equivalence, the closeness or sameness of form has been made tertium comparationis. Several phraseological units that are the same or similar in form but different in meaning in English and Slovene are analysed in the paper. Some of these pairs of idioms show certain common features, such as comparison, emotion, spoken or written communication. Phraseological false friends are illustrated by examples and similarities and differences between the idiom in English and the phraseological false friend in Slovene are commented upon. Since phraseological as well as lexical false friends represent a great problem in communication, translation and lexicographic treatment, it is necessary to first raise awareness of the lexical traps into which non-native speakers of English as well as any other language may easily fall, regardless of their level of linguistic knowledge. It is, therefore, essential to find and treat these pairs of idioms appropriately and acquaint learners with them by including them in course books, in bilingual, general and especially phraseological dictionaries.
    [Show full text]
  • TKT Module: 1 Describing Language: Lexis – Teacher's Notes
    TKT Module: 1 Describing language: Lexis – Teacher’s Notes Description This activity covers the TKT Module 1 Part 1 syllabus area of concepts and terminology for describing language relating to lexis. Teachers consider the various aspects of lexis that learners may need to know and do a practice activity. Time required: 60 minutes Materials . Participant’s Worksheet 1 (one per pair) required: . Participant’s Worksheet 2 (one for each participant) . Participant’s Worksheet 3 (cut into strips, 1 set per pair) . Participant’s Worksheet 4 (cut into strips, 1 set per pair) . Sample Task (one for each participant) Aims: . To provide an opportunity for participants to discuss lexis: types of meaning, word formation, word groupings and register. To provide practice in completing tasks in which lexis is the testing focus. Procedure 1. Before the session, copy Participant’s worksheets 3 and 4 so there is one copy of each for every pair of participants. Cut each worksheet into strips and keep the sets of strips together. 2. (10 minutes) On the board write the following letters arranged in a circle on the board. A C R V B A U Y O L. Ask participants to arrange the letters so make a word which has something to do with language learning (vocabulary). Write this on the board. Elicit another word which means the same as vocabulary (lexis) and write this on the board. 3. Give each pair a copy of Participant’s worksheet 1. First, ask them to fold over the box at the very bottom of the worksheet (they don’t need to look at this until Step 7).
    [Show full text]
  • The Problem of False Friends in Learner Language: Evidence from Two Learner Corpora 1
    The problem of false friends in learner language: Evidence from two learner corpora 1. False friends are a real problem for language learners. Even so, little research has been done on the identification of the difficulties learners have when it comes to the use of these words. The aim of this paper is to analyze false friends in the interlanguage of Spanish learners of English. Two learner corpora: the Santiago University Learner of English Corpus (SULEC) and the International Corpus of Learner English (ICLE), representative of the students´ interlanguage, will allow us to identify the extent of the problem. The conclusions of this study will provide new insights into the linguistic and the communication problems derived from a misuse of these lexical items. Key words: false friends, learner corpora, language learning. Los falsos amigos constituyen un problema real para los estudiantes de lenguas. Aún así, se ha invertido poco tiempo en la identificación de las dificultades que los estudiantes presentan al utilizar estas palabras. Esta comunicación tiene como objetivo analizar los falsos amigos en la interlengua de los estudiantes españoles de inglés. Dos corpus de aprendices: el Santiago University Learner of English Corpus (SULEC) y el International Corpus of Learner English (ICLE), representativos de la interlengua de los estudiantes, nos permitirán identificar la dimensión del problema. Las conclusiones extraídas nos ayudarán a entender los problemas lingüísticos y de comunicación derivados de un mal uso de estos elementos léxicos. Palabras clave: falsos amigos, corpus de aprendices, aprendizaje de lenguas. 1. INTRODUCTION False friends have become a real problem for language learners.
    [Show full text]
  • Towards a Typological Classification of False Friends (Spanish-English)
    RESLA 19 (2006), 29-39 TOWARDS A TYPOLOGICAL CLASSIFICATION OF FALSE FRIENDS (SPANISH-ENGLISH) RUBÉN CHACÓN BELTRÁN UNED ABSTRACT. Cognate words in two or more languages have a common origin because of their diachronic relationship and, as a result, they share some sort of formal and/or semantic affinity. Cognate words can facilitate the foreign language learning process; they have similar meanings and, therefore, they can support the acquisition and/or learning of a non-native language. However, these words can also have a deceptive meaning as a result of semantic change and dissimilar development in two languages, i.e., they may be deceptive cognate words or false friends. False friends are especially problematic for language learners as they tend to overgeneralize and assume they know the meaning of these words, which are actually misleading. In this paper a taxonomy of cognate words containing six different types is put forward. The classification is based on semantic and formal criteria (morphological and phonological) and aims at providing a common framework for the analysis of cognate words which will help in the development of specific teaching and learning strategies. KEYWORDS. Vocabulary learning, crosslinguistic influence, false friends, cognate words. RESUMEN. Las palabras cognadas en dos o más lenguas son aquellas que tienen un origen común dada su relación desde un punto de vista diacrónico y, como conse- cuencia de ello, comparten afinidades formales y/o semánticas. Los cognados pueden favorecer el proceso de aprendizaje de una lengua extranjera, pues tienen significados similares y así fomentan la adquisición y/o el aprendizaje de la lengua no nativa.
    [Show full text]
  • False Friends
    ZÁPADOČESKÁ UNIVERZITA V PLZNI FAKULTA PEDAGOGICKÁ KATEDRA ANGLICKÉHO JAZYKA MEZIJAZYKOVÁ HOMONYMIE: NEPRAVÉ EKVIVALENTY BAKALÁŘSKÁ PRÁCE Lucie Šefčíková Anglický jazyk se zaměřením na vzdělání Vedoucí práce: PhDr. Naděžda Stašková, PhD. Plzeň, 2016 UNIVERSITY OF WEST BOHEMIA FACULTY OF EDUCATION DEPARTMENT OF ENGLISH INTERLANGUAGE HOMONYMY: FALSE FRIENDS UNDERGRADUATE THESIS Lucie Šefčíková English Language in Education Supervisor: PhDr. Naděžda Stašková, PhD. Pilsen, 2016 Prohlašuji, že jsem bakalářskou práci vypracovala samostatně s použitím uvedené literatury a zdrojů informací. Plzeň, 20. dubna 2016 .......................................................................... vlastnoruční podpis ACKNOWLEDGMENTS I would like to thank my supervisor PhDr. Naděžda Stašková, PhD. for a patient guidance. My special words of thank go to my family and friends for their support and help, and finally to my Muse which has often visited me. ABSTRACT Šefčíková, Lucie. University of West Bohemia. April, 2016. Interlanguage homonymy. False friends. Supervisor: PhDr. Naděžda Stašková, PhD. The object of this undergraduate thesis is the linguistic phenomenon of false friends. It consists of two main parts, theoretical and practical one. The theoretical part of this study deals with the treacherous words from the synchronic point of view providing the reader a brief overview of false friends and its classification. Secondly, the diachronic point of view is discussed and thus in relation to the greatest English playwright William Shakespeare. It demonstrates the change of meanings of words over time. The practical part focuses on the general awareness of false friends. Four categories including Czech students of English as well as of other subjects and people working in the Czech Republic as well as abroad were examined. The conducted research was realized by means of questionnaires which firstly aimed at the personal information of respondents.
    [Show full text]
  • False Friends in Advanced Learners of English: the Effect of Task Type and Mode
    False friends in advanced learners of English: the effect of task type and mode Supervisor: Dr. Elisabet Pladevall Ballester Gemma Solé Alonso TFG Estudis Anglesos Universitat Autònoma de Barcelona June 2017 Table of contents Acknowledgements ...................................................................................................... - 1 - Abstract ......................................................................................................................... - 1 - 1 Introduction ........................................................................................................... - 1 - 2 Acquisition of False friends: State of the art and literature review ....................... - 4 - 2.1 Previous research on the acquisition of false friends ..................................... - 4 - 2.2 Oral and written modes ................................................................................ - 10 - 2.3 Production and comprehension .................................................................... - 12 - 3 Methodology ....................................................................................................... - 15 - 3.1 Participants ................................................................................................... - 15 - 3.2 Tasks ............................................................................................................ - 15 - 4 Results and data analysis ..................................................................................... - 17 - 4.1 Oral and written
    [Show full text]
  • FALSE FRIENDS”: a REVIEW of SOME USEFUL RESOURCES Mª Luisa Roca Varela University of Santiago De Compostela
    TEACHING AND LEARNING “FALSE FRIENDS”: A REVIEW OF SOME USEFUL RESOURCES Mª Luisa Roca Varela University of Santiago de Compostela Resumen Denominamos falsos amigos a palabras similares en dos o más lenguas que poseen significado/s diferentes ( library “biblioteca” y librería “tienda de libros”, en inglés y en español respectivamente). Desde el punto de vista de la enseñanza y el aprendizaje del inglés, los falsos amigos son importantes porque con frecuencia nos llevan a errores en la expresión, y malentendidos a nivel de comprensión (e.g. I am constipated; it has been a fastidious speech or you should eat food without preservatives). Este artículo pretende dar a conocer una serie de herramientas útiles (diccionarios, recursos en línea, glosarios y blogs) para el estudio de los falsos amigos. Los recursos que aquí se presentan contribuirán a comprender la complejidad del fenómeno de los falsos amigos y aportarán la información necesaria para tratar este tema en el aula de inglés. Palabras Clave : falsos amigos, aprendizaje y enseñanza del inglés, diccionarios, recursos en línea Abstract False friends are words in two languages that are similar in form but different in meaning (e.g. English library “place for reading and borrowing books” vs Spanish librería “bookshop”). From the point of view of EFL teaching and learning, false friends are important because they lead us to errors in L2 production and comprehension (e.g. I am very constipated; it has been a fastidious speech and you should eat food without preservatives). This paper intends to present some useful tools (dictionaries, online resources, glossaries and blogs) for the teaching and learning of false friends.
    [Show full text]
  • Department of English and American Studies
    Masaryk University Faculty of Arts Department of English and American Studies English Language and Literature Lenka Tycová Towards a New False Friends Dictionary Bachelor‟s Diploma Thesis Supervisor: Ing. Mgr. Jiří Rambousek 2012 I declare that I have worked on this thesis independently, using only the primary and secondary sources listed in the bibliography. …………………………………………… Author‟s signature Acknowledgement: I would like to thank my supervisor Ing. Mgr. Jiří Rambousek for his patience, support, valuable advice and kind guidance. Table of Contents 1.0 Introduction .......................................................................................................................... 6 1.1 Motivation ........................................................................................................................ 6 1.2 Objective ........................................................................................................................... 7 2.0 Defining the Concept of “False Friends” ............................................................................. 8 2.1 Putting it Right: “False Friend” or “False Cognate”? ..................................................... 10 3.0 Classification of False Friends: Categorizing the Uncategorizable? ................................. 13 4.0 Approaches to the Treatment of False Friends ................................................................... 23 5.0 Towards a New Czech-English-English-Czech False Friends Dictionary ......................... 28 5.1 Scope .............................................................................................................................
    [Show full text]
  • Matoucí Dvojice Slov
    Západočeská univerzita v Plzni Fakulta pedagogická Katedra anglického jazyka Bakalářská práce MATOUCÍ DVOJICE SLOV Kateřina Bílková Plzeň 2018 University of West Bohemia Faculty of Education Department of English Undergraduate Thesis CONFUSING WORD PAIRS Kateřina Bílková Plzeň 2018 Prohlašuji, že jsem práci vypracovala samostatně s použitím uvedené literatury a zdrojů informací. V Plzni dne 27. dubna 2018 ……………………………. Jméno Příjmení ACKNOWLEDGMENTS I would like to express my gratitude to PhDr. Naděžda Stašková, Ph.D. for her professional and patient guidance and time. I would also like to thank my family, friends and especially my partner for their support and help. ABSTRACT Bílková, Kateřina. University of West Bohemia. April, 2018. Confusing Word Pairs. Supervisor: PhDr. Naděžda Stašková, Ph.D. This undergraduate thesis is concerned with the topic of confusing word pairs which is relevant not only for learners of English language but also for native speakers. It consists of two main parts, theoretical and practical one. The theoretical background provides the reader cohesive information concerning lexical meaning and paradigmatic relations with a special focus on homonymy. Four main groups of homonyms are selected and further elaborated, specifically grammatical homonyms, homophones, paronyms and interlingual homonyms, also called false friends. However, terms such as absolute homonymy, partial homonymy, proper homonymy, oronymy and homography are included as well because of their close relation to the subject matter. In the practical part of the thesis, concrete examples of commonly confused word pairs are presented. The analysis is based on the explanation of the lexical meaning of each expression and on finding particular features within the selected group of homonyms.
    [Show full text]