A Comprehensive Classification Model for Sinhala Compound Nouns

RANDULA INTRODUCTION RESEARCH PROBLEM & PURPOSE RESEARCH METHODOLOGY FINDINGS & DISCUSSION CONCLUSIONS & TAKEAWAYS  Sinhala [sɪŋhələ | සිංහල] is an Indo-Aryan language primarily spoken by the Sinhalese people of Sri Lanka, who make up the largest ethnic group on the island, numbering about 16 million. Sinhala is also spoken as the first language by other ethnic groups in Sri Lanka, totalling about 2 million speakers. It is written using the Sinhala script, which is a Brahmic script closely related to the Grantha script of South India. Sinhala has been enriched from its prolonged contact with Sanskrit, Pali, Tamil, Arabic, Portuguese, Dutch, Malay, and English. Divehi, the language spoken in Maldives is an off-shoot of Sinhala. 4 • A compound is a new lexeme formed by combining two or more stems, each capable of functioning independently in other contexts (Bauer, 2001: 695). Compounding is the concatenative process of forming compounds by ordering morphemes one after another (Hapselmath & Sims, 2010: 34). In Sinhala, there are three main types of compounds as compound nouns, compound adjectives, and compound verbs. The focus of this study is only on compound nouns in Sinhala. • This study intends to address the limitations of the current classification model of Sinhala compound nouns by proposing a more comprehensive model. The traditional classification model, introduced in the traditional grammar, Sidat Sangara (13th C. AD) and later adopted by school grammars, has faced criticism from modern linguists for its descriptive inadequacy and inability to capture the distinctive features of Sinhala compound nouns (Silva, 1972: 79; Disanayaka, 2010: 65). 5 • The problem addressed by this study is the inadequacy of the current classification system for compound words in the Sinhala language. Despite its historical significance, the existing classification model introduced by the Sidat Sangara roughly imitating the Sanskrit model, is overly simplistic and fails to capture the complexity and richness of Sinhala compounds. • The purpose of this study is to propose a comprehensive classification model for Sinhala compound nouns that encompasses orthographic, phonological, morphological, typological, semantic, etymological, and historical dimensions, providing a more accurate portrayal of Sinhala compound nouns. 6 • Compounding in Sinhala remains an understudied area owing to the skeptical attitudes of prominent structural linguists: “... සමාස යනුවෙන් අප උගන්නා දැය නිකරුවේ කාලය අපවේ යැවීවමහි ලා කදිම උපායකි. ‘සුදු මල’ සමාසයකි. සමාස නුවූ කල එය සටිනුවේ වකවසේ ද? ඒේ, ‘සුදු මල’ යනුවෙනි. එවසේ නම්, සමාස වීම යනුවෙන් කුමක් හැවෙයි ද?...” (ද සල්ො, 1972: 79) “…What we study as compounding is a wonderful exercise in futility. For instance, ‘sudu mɑlə’ (literally, ‘white flower’) is considered a compound. But, if it were not compounded, how would it be? It would still be ‘sudu mɑlə’ (literally, ‘white flower’). So, then, what does compounding really mean?...” (de Silva, 1972: 79) [Translated by the presenter.] • Is this really the case? It is worth exploring. 7 This study employs a desk research approach to comprehensively investigate the etymology, morphology, and semantics of Sinhala compound nouns within the domain of contemporary descriptive linguistics. 1. Data Collection Methods  The foundation of the study lies in an extensive review of literature, linguistic references, and grammatical sources pertaining to Sinhala compound words. Secondary data from these sources were analysed to pinpoint limitations in the current classification system.  Primary data collection involved observing a convenient sample of native Sinhala speakers to capture compound nouns as used in everyday speech. This approach provided invaluable perspectives not fully captured in traditional grammatical resources. 2. Data Analysis Methods  The study conducted a thorough examination of the etymological origins of Sinhala compound nouns, tracing their linguistic evolution and exploring influences from various parent and contact languages.  Following the etymological analysis, a detailed investigation into the phonological and morphological structures was undertaken, with a focus on syntactic categories and structures. The semantic analysis centred on criteria such as case relationships, dependency, and headedness. 8 • The Sidat Sangara written in the 13th century AD presents a five-fold classification model for compound nouns in Sinhala. All successive grammars have adapted this classification forcing new examples into the five classes proposed in the Sidat Sangara. • The classification in Sidat Sangara is not an authentic description of the Sinhala language, but a simplification of the six compound classes in Sanskrit into five overlapping classes. 1. 2. 3. 4. 5. 10 A few defining features of Sinhala compounds are as follows (Karunatillake, 1995: 246-249): • Morphological features: • word-final inflection • right-headedness • internal integrity • Syntactic features: • single construction • positional mobility • Semantic features: • single meaning unit (mostly) • centricity/ headedness • endocentric/ compositional – head + determinant(s) OR head1 + head2 (+head 3) • exocentric/ epithetic – external head 11 • Generative features: • Every compound [surface structure] derives from an underlying defining clause [deep structure] that determines the syntactic relations among compound components. Eg1: sudu mɑlə [N+N] – Adjectival Compound Surface Structure: Deep Structure: sudu mɑl-ə mɑl-ə sudu pɑ:ʈə white flower-DEF.SG flower-DEF.SG white colour ‘white flower’ ‘The flower is white in colour.’ Eg2: kiri bɑtə [N+N] – Subordinative/ Case Compound Surface Structure: Deep Structure: kiri bɑt-ə ejɑ: kir[i]-en bɑtɑk milk rice-DEF.SG s/he milk-INST rice-IND.SG ‘coconut-milk rice’ ‘S/he cooks a rice with milk.’ we: be-3.SG.PRS. ujə-nə-wɑ cook-PRS-FNT. 12 1.1 Open/ Spaced Compounds In open compounds, the component lexemes are written separately with a space in their internal boundary, but together they refer to a single referent. E.g.: eɭu mɑs (‘goat meat’ - mutton), rɑtu gonɑ: (red bull), kiri ko:pi (milk coffee) 1.2 Closed/ Solid Compounds In closed compounds, the component lexemes are combined into one single form without internal spaces. These might or might not involve euphonic combinations in internal boundaries. E.g.: kiribɑt (milk rice), nowijɑt (‘non-scholarly’ ill-educated), niʃɑ:kərə (‘night causer’ - moon) 1.3 Hyphenated Compounds In hyphenated compounds, the lexemes are linked by a hyphen to show that they work together as a single unit. Only coordinative/ copulative compounds are written like this. E.g.: guru-deguru (teacher-parent), swɑ:mi-se:wəkə (master-servent), iɖɑm-kɑɖɑm (lands-plots) 14 - Chandralal, 2010: 89 Grammatically, a single word is a single unit of inflection. It does not necessarily need to appear as a single orthographic word, as orthographic unity is not a defining feature of internal integrity. Therefore, compound nouns may or may not have internal boundaries. 15 Sinhala compounds can be broadly classified into two categories based on whether the internal lexeme boundary is phonologically transparent or opaque. 2.1 Conjoined Compounds Conjoined compounds involve a euphonic combination process that phonologically modifies the boundary between the two lexemes, creating a seamless blend. All conjoined compounds are essentially closed/ solid compounds as their boundaries are merged. E.g.: nilupul (blue lotus), bɑ:ʃɑ:dhjəjənəjə (‘language study’ linguistics), rɑnɑmbərə (golden sky), liggɑlə (hearth stone), nawo:dəjə (new down), butsɑrəɳə (‘Buddha refugee’ – refugee of the Buddha) 2.2 Disjoined Compounds Disjoined compounds, do not involve any euphonic combination process. The two lexemes retain their original forms without phonological modifications at the boundary, meaning they are simply placed together without blending sounds. All disjoined compounds are essentially open compounds as their boundaries are retained. E.g.: nil upul (blue lotus), bɑ:ʃɑ: ɑdhjəjənəjə (‘language study’ linguistics), rɑn ɑmbərə (golden sky), ɭip gɑlə (hearth stone), nɑwə udəjə (new down), budu sɑrəɳə (‘Buddha refugee’ – refugee of Buddha) 16 Sinhala compound nouns can be categorised into ten morphological categories considering the word classes of the lexemes involved in them: 3.1 N(Adj.) + Noun: sudu mɑl (white flowers), uɖərɑtə (up country), mɑhɑtə minisɑ (fat man) 3.2 N(Arg.) + Noun: minis sɑ:mɑɈəjə (human society), bɑlu wɑligəyə, mɑs mɑ:lu (meat [and] fish) 3.3 Pr. Participle + Noun: lijənə me:səjə (writing table), bonə lində (drinking well) 3.4 Ps. Participle + Noun: kædiccə putuwə (broken chair), bindunu bɑndun (broken vessels) 3.5 Noun + Pr. Participle: jɑturu lijənəjə (typewriter), durə dɑknajə (‘distance seer’ - telescope) 3.6 Gerund + Noun: lijum kɑwərəjə (‘letter cover’ - envelope), kæwili me:səjə (sweets table) 3.7 Noun + Gerund: rɑsə kæwili (‘tasty eats’ - sweets), bɑɖu wikini:mə (selling goods) 3.8 Gerund + Gerund: aenum bænum (hits [and] scolds), kæɖum bindum (breakings [and] cracks) 3.9 Particle + Noun: jɑ:waɈi:wə (‘extent life’ - lifelong), tɑ:wəkɑ:likə (temporary time) 17 Sinhala compounds can be broadly classified into two categories based on the derivational origin of their heads: those with a verbal-noun head and those with a pure-noun head. 4.1 Deverbal/ Synthetic Compounds Synthetic compounds are composed of two lexemes, where the head lexeme is derived from a verb, and the non-head is interpreted as an argument (usually the object) of that verb (Lieber, 2015: 46). E.g.: gɑtkɑru (‘book doer’ – writer), gɑs mɑdinnɑ (‘tree tapper’ – toddy tapper), mi:pɑl (‘land ruler’ – king), dɑlu neɭɑnni (‘bud plucker’ - tea leaf picker) 4.2 Denominal/ Primary Compounds Denominal compounds are made up of two lexemes, which may be nouns, adjectives, or verbs; the second lexeme is typically derived from a noun (Lieber, 2015: 46). E.g.: gɑtwɑhɑrə (‘book language’ – formal lang.), gɑs mɑnɖijə (tree grove), mi:dum (‘land smoke’ – mist), dɑlu murəjə (‘bud shift’ - tea leaf picking shift) 18 Sinhala compounds are classified into two broad classes based on their semantic centricity – where the semantic head lies, inside or outside the construction. 5.1 Endocentric Compounds In endocentric compounds, the referent of the compound is always the same as the referent of its head lexeme (Lieber, 2017: 48). E.g.: pol malə (coconut fower), thihini nuwənə (sharp wisdom), balu waligəyə (dog’s tail), priyankərə (‘pleasant causer’ - pleasant), sorə bijə (‘thives fear’), gunə daham (values virtues) 5.2 Exocentric Compounds Compounds may be termed exocentric when the referent of the compound as a whole is not the referent of the head lexeme (Lieber, 2017: 48). E.g.: kehel mala (‘banana flower’ – nonsense), hiwal nuwana (‘fox wisdom’ - cunningness), koti waligaja (‘tiger tail’ – a never ending problem), nisa:kara (‘night causer’ - moon) 19 6.1 Determinative Compounds A determinative compound is a type of compound word in which one lexeme (the determinant) modifies or specifies the meaning of the other lexeme (the head), giving the compound a more specific meaning than either lexeme alone. These are of two types depending on the syntactic relationship between the components: 6.1.1 Attributive Compounds (N.Adj + N) If the determinant attributively modifies the head of a compound like an adjective does, it is an attributive compound (Lieber, 2017: 222). The head and determinant have a relationship similar to that of the subject and its complement in a copula-verbed sentence. 6.1.1.1 Adjectival Compounds (N.Adjective + N) E.g.: usa minisa (tall man), maha: purshaja: (great man), loku aiyya (big brother), lassana mal (beautiful flowers), katho:lika pijatuma (‘Catholic father’ – Catho. priest) 6.1.1.2 Numeral Compounds (N.Numeral + N) E.g.: tun lo:kaya (three worlds), sadewlo: (seven heavens), sat samudura (seven oceans), tun tis peja (‘thiry three hours’ – more than half a day), pan sil (five precepts) 20 6.1.1.3 Particle Compounds (Particle + N) Particle compounds include an indeclinable particle the determinant while having a noun as the head. Unlike in Sanskrit and Pali, particle compounds in Sinhala always result in compound nouns as their head lexeme is always a noun. E.g.: ja:waɉi:wa (‘extent life’ – lifelong), ta:waka:lika (cotemporary time - temporary), saho:dara (‘shared womb’ - sibling), jathalaba (proper profit), jatha:ru:pa (proper image), sahawæsi (‘shared habitation’ - cohabitants), sahasambandha (inter-connected), nominis (non-human), nowijath (‘non-scholar’ - illiterate) Although traditional grammars blindly follow Sanskrit and include prefixed nouns under particle compounds, it has been understood to be theoretically incorrect (Cumaratunga, 1939: 329-330). Hence, in this study, such words are not considered. 21 ATT. COMPOUNDS – DEEP STRUCTURE • 1: Adjectival Compound [N + N] Surface Structure: lɑssənə mɑl-ə beautiful flower-DEF.SG ‘the beautiful flower’ Deep Structure: mɑl-ə lɑssənə flower-DEF.SG beautiful ‘The flower is beautiful.’ wei be-PRS. Deep Structure: sɑmudur-u sɑt-ɑk sea-PL seven-IND.SG ‘There are seven seas.’ tije-nə-wɑ exist-PRS-FNT. Deep Structure: e:kɑ: minih-ek s/he human-IND-SG ‘S/he is not a human’ wei be-PRS. • 2: Numeral Compound [N + N] Surface Structure: sɑt sɑmudur-ə milk rice-DEF.SG ‘the seven seas’ • 3: Particle Compound [Part + N] Surface Structure: no-minih-ek NEG-human-IND.SG ‘a non-human’ no not 22 6.1.2 Subordinative/ Case Compounds (N + N) A subordinative compound is a type of compound where the determinant modifies the head, typically specifying the relationship between them, usually in terms of grammatical cases (Lieber, 2017: 230). Hence the determinant is subordinate to the head and interpreted as an argument of a deep structure sentence with a lexical verb. 6.1.2.1 Accusative C. Compound: wedakaru (‘work dort’ - servent), nasthika:ra (waste doer), sampathda:yaka (resource giver) 6.1.2.2 Agentive C. Compound: paraputu (others fed - parasite), ma:nawakruthi (human done), bambadath (Brahma given), by humans) 6.1.2.3 Instrumental C. Compound: rankam (gold work), kadupa:ra (stike sword), atkam (handicrafts), diyabath (water rice) 6.1.2.4 Dative C. Compound: pinbath (free rice), sama:ja se:waya (social work), janasawiya (people strength), kurulu pa:luwa (bird feeding) 6.1.2.5 Ablative C. Compound: sorabiya (fear of theaves), widesha a:da:ra (foreign aid), mi:dum (land smoke), rataha:l (country rice) 6.1.2.6 Genitive C. Compound: kanta wimuktija (women liberation), sanga parapura (monk generation), guru widyalaya (teacher college) 6.1.2.7 Locative C. Compound: poltel (coconut oil), bat dankuda (scorched rice), aenda redda (bed sheet), taembili tel (king coconut oil) 23 SUBORD. COMPOUND – DEEP STRUCTURE • 1: Ablative Case Compound [N + N] Surface Structure: wideʃə ɑ:dhɑ:rə foreign aid ‘foreign aid’ Deep Structure: ɑpi-ʈə wide:ʃə-j-en 1.PL-DAT foreign-DEF.SG-ABL ‘We get foreign aid.’ ɑ:dhɑ:rə aid læbe-nə-wɑ get-PRS-FNT • 2: Dative Case Compound [N + N] Surface Structure: sɑmɑ:jə se:wə-jə society service-DEF.SG ‘the seven seas’ Deep Structure: ejɑ sɑmɑ:ɟə-jə-ʈə səwəjə 3.SG society-DEF.SG-DAT service ‘S/he does (some) service to the society.’ kərə-nə-wɑ do-PRS-FNT • 3: Particle Compound [Part + N] Surface Structure: ændə redd-ə bed cloth-DEF.SG ‘the bedsheet’ Deep Structure: ænd-e: redd[i]-ɑk bed-LOC cloth-IND-SG ‘There is a sheet on the bed.’ tije-nə-wɑ exist-PRS-FNT 24 6.2 Copulative Compounds A copulative compound is a type of compound in which both lexemes of the compound have equal importance or status, and neither lexeme modifies the other (Lieber, 2015: 47). Instead, both lexemes of the compound are typically on equal footing, and the compound as a whole represents the sum or combination of the two lexemes. 6.2.1 Appositional Copulative Compounds In an appositional copulative compound, all components refer to the same entity. E.g.: le:kaka-prakashaka (writer-publisher), shishyaguru (teacher-student), widya:-shilpa-ta:kshanagna (science, craft and technologist) 6.2.2 Coordinative Copulative Compounds In a coordinative copulative compound, each component refers to a separate entity. E.g.: gi: naetum (songs [and] dance), ruk liya (trees [and] wines), pat pot (papers and books), du:puthun (daughters [and] sons), palu pala mal (buds, fruits [and] flowers), sanda hiru tharu (moon, sun [and] stars), aeala dola gang (cannels, brooks [and] rivers) 25 All determinative compounds (attributive and subordinative) can be broadly classified into two categories based on whether the inflections of the first component are elided or retained in compounding. 7.1 Elliptical Compounds Most compounds are elliptical compounds, where the case denoting suffixes of the first lexeme are elided in compounding. E.g.: badasa: (‘stomach fire’ – hunger), temas da:naya (three-month alms giving), eas weda: (eye doctor), atbehet (‘hand medicine’ – home remedy) 7.2 Non-elliptical Compounds Non-elliptical compounds retain the case denoting suffixes of the first lexeme even when they are compounded with another lexeme. E.g.: bade ama:ruwa (stomachache), hatdohe: da:ne (seventh day’s alms giving), rate veda (‘village’s doctor’ – someone pokes his nose to everything), ataeweesi (proximity liver – apprentice) 26 A binomial compound is a coordinated pair of lexemes of the same word class with some semantic relation (Kopaczyk, & Sauer, 2017: 3). This is a special category of copulative compounds that always contain only two lexemes – thus the name binominal compounds. These are classified semantically and phonemically. 8.1 Semantic Classification 8.1.1 Synonymous Binominals: leda ro:ga (diseases illnesses), mas ma:nsa (meat meat) 8.1.2 Antonymous Binominals: pin paw (merits sins), honda naraka (good bad) 8.1.3 Hyponymous Binominals: daru daeri (boys girls), nalu nili (actors actresses) 8.1.4 Dyadic Binominals: ambu saemi (husband wife), guru sisu (teacher student) 8.1.5 Collocative Binominals: nadu haba (cases and quarrels), ge: dora (house compound) 8.2 Phonemic Classification 8.2.1 Rhyming Binominals: dan pahan (alms and lamps), kea:m bi:m (food and beverages) 8.2.2. Reduplicative Binominals: aeta beta (bones - ?), andu pandu (claws – ? 8.2.3. Onomatopoeic Binominals: kasukusu (chitter chatter ), hutapata (double trouble) 27 This classification is based on the histiorical dating of compounds and the etymological categories (adopted – tatsama vs. adapted – tadbhawa) of their components. 9.1 Classical Compounds 9.1.1 Adopted Compounds: diva:kara (sun), de:wadaththa (god gievn), dhrama deshanaya (sermon) 9.1.2 Adapted Compounds: diwayuru (sun), dewdath (god given), daham desuma (sermon) 9.2. Neo-Classical Compounds 9.2.1 Adopted Comps: desapa:lana widja:wa (political sc.), pariganaka jantraja (computer machine) 9.2.2 Adapted Comps: pa:sal poth (schoolbooks), dura daknaja (telscope), pa:paediya (bicycle) 9.3 Hybrid Compounds/ Loan Blends 9.4.1 Adopted + Adapted Compounds: bhu:mi thel (kerosine), candra madala (lunar halo) 9.4.2 Adapted + Adopted Compounds: pa:sal bhu:mija (school land), aes pari:kshawa (eye check) 9.4.3. Adopted1 + Adopted2 Compounds: lo:ka benkuwa (word bank), kalikan kakula (trouser leg) 9.4 Colloquial Compounds: gam mulae:ni (village headman), bo:kku kata (culvert opening) 28 - Chandralal, 2010: 44 29 Tautological or pleonastic compounds refer to compounds in which both lexemes essentially mean the same thing, leading to redundancy in meaning. In such compounds, the first lexeme is typically a lesser-known one of which the meaning is elaborated by the second lexeme which is a rather wellknown lexeme meaning the same. So, in such compounds, the second lexeme glosses the first. The first lexeme is lesser known because it is usually archaic or foreign: 10.1 Archaic + Modern Compounds E.g.: [sadath] thala:wila (pond pond), [thiri] wa:nagala (stone stone), [mul] kirigala (rock rock), wa:saldora (door door), wana arana (forest forest), wili lejjawa (shame shame) 10.2. Foreign + Native Compounds E.g.: board le:lla (bord board), blade thalaja (blade blade), isasan pola (station station), light eliya (light light), company sama:gama (company company), cathch aelli:ma (catch) 10.3. Adapted + Adopted Compounds E.g.: badu bhanda (goods), awi ajudha (weapon), puda po:ja (offerings), raeki raksha (jobs) 30 - Chandralal, 2010: 87 31 • This study has identified that while Sinhala compounds exhibit ten distinct types of morphological structures, these compounds can be broadly classified into two typological categories based on the dependency relations in their deep structure clauses as determinative and copulative compounds. • Additionally, several dichotomies such as open-closed-hyphenated, endocentric-exocentric, denominal-deverbal, elliptical-non-elliptical, and conjoined-disjoined compounds are observed. • Adjectival, case, and particle compounds, traditionally classified separately, are recognised as subtypes of determinative compounds as all their determinant lexemes perform the common function of modifying the head lexeme in some manner. • Binominals, not recognised by previous researchers as compound nouns, are identified as a major subtype of copulative compounds. This study acknowledges the significance of tautological compounds as a result of language change. • This refined model provides a comprehensive framework for understanding the linguistic complexity of Sinhala compound nouns. Future research could further explore the practical applications and implications of this classification model in educational and computational linguistics contexts. 32 Bauer, L. (2001). ‘Compounding’ in Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible (eds.), Language Universals and Language Typology, Vol. 1, 695–707. Berlin and New York: de Gruyter. Bauer, L. (2004). A Glossary of Morphology. Edinburgh: Edinburgh University Press. Bauer, L., Lieber, R., & Plag, I. (2014). The Oxford Reference Guide to English Morphology. London: Oxford University Press. Chandralal, D. (2010). London Oriental and African Language Library – 15: Sinhala. Amsterdam: John Benjamins Publishing. Crystal, D. (2008). A Dictionary of Linguistics and Phonetics (6th ed.). Oxford: Blackwell Co. Cumaratunga, M. (1938). Vyākaraṇa Vivaraṇaya [Elucidation of Grammar]. Colombo: M. D. Gunasena & Co. de Silva, M. W. S. (1972). Bhāshā Vishayaka Lipi [Papers on Language]. Colombo: Lakehouse Bookshop. Dharmarama, Ven. R., (ed.) (1902). Sidat Sag̃arā Vistrara Sannaya [A Descriptive Commentary on the Sidat Sangara]. Peliyagoda: Vidyalankara Press. Disanayaka, J. B. (2010). ‘An Invitation to Sinhala Language Teacher from the Global Village’ in Samantha Herath, Percy Jayamanne, Praneeth Abeysundara, and Thimbiriyagama Bandara (eds.), Sāhitya: Special Volume 2010. Colombo: Department of Cultural Affairs. 33 Gamlath. S. (2012). Nawa Sinhala Vyākaraṇaya – 5: Sinhala Samasa [New Sinhala Grammar – 5: Sinhala Compounds]. Colombo: S. Godage & Brothers. Jayawardhana, T. (1994). ‘Compound Words in Sinhala’ in D. M. Wickremasinghe & H. Weerakoon (eds.), Wāgvidyā – 5. Kelaniya: Dept of Linguistics, Kelaniya. Karunatillake, W. S. (1995). Sinhala Bhāshā Vyākaraṇaya [A Grammar of the Sinhala Language]. Colombo: M. D. Gunasena & Co. Kopaczyk, J., & Sauer, H. (Eds.). (2017). Binomials in the History of English: Fixed and Flexible. Cambridge: Cambridge University Press. Lieber, R. (2017). Introducing Morphology (2nd ed.). Cambridge: Cambridge University Press. Vitharana, W. (2013). Viyaraṇa Visitura [The Beauty of Grammar]. Colombo: S. Godage Brothers. Wijeratne, W. M. (1991). ‘Compound Nouns’ in W. M. Wijeratne & Asoka Premaratne (eds.), Wāgvidyā – 4.Kelaniya: Department of Linguistics, Kelaniya. 34