Reduplication occupies a middle ground between concatenative and non-concatenative morphology. Reduplication involves copying phonological material from a base (a root or stem) to create a reduplicant (the repeated element) and affixing it to the base to create a new stem. It is still a form of affixation, since the resulting stem is easily segmented. But unlike typical concatenative morphology, in reduplication, the the form of the affix – the reduplicant – depends on the form of the base.1
Although Western European languages2 do not tend to use reduplication productively, it is a very common type of morphology in the world's languages. In fact, 85% of languages sampled have productive reduplication of some sort, and only 15% lack it.3 Among pidgins and creoles, the proportion is about the same: 86.8% of the languages in APiCS,4 the pidgin and creole database, have reduplication.
Formal properties of reduplication
Reduplication usually seems to target (a portion of) the morphological root. Only rarely are affixes targeted to the exclusion of the root. An example of affix-targeting reduplication can be found in Amele, where reduplication of the object marker is used to express iterative aspect. When no object marker is present, the root is taken as the base.
Amele (Trans New Guinea, Madang)5
a. hawa-du 'ignore it' → hawa-du-du 'keep ignoring it'
b. gobil-du 'stir it' → gobil-du-du 'stir and stir it'
c. guduc-du 'run it' → guduc-du-du 'keep running it'
d. qu 'hit' → qu-qu 'keep hitting'
e. ji 'eat' → ji-ji 'keep eating'
f. budu-eɁ 'to thud' → budu-budu-eɁ 'to thud repeatedly'
g. g͡batan-eɁ 'to split' → g͡batan-g͡batan-eɁ 'to split repeatedly'
Full and partial reduplication
Reduplication can be full or partial. Full reduplication is when the reduplicant is equal to the whole base. Full reduplication occurs in many languages. Here an example is drawn from Sundic, where the plural of kərá 'monkey' is formed by repeating the base in its entirety:
Sundic (Western Malayo-Polynesian)6
kərá 'monkey' → kərá-kərá 'monkeys'
Partial reduplication, by contrast, occurs when the reduplicant is equal only to part of the base. For example, in Agta, the plural of takki 'leg' is formed by reduplicating only the portion tak of the base takki.
Agta (Western Malayo-Polynesian)
takki ‘leg’ → tak-takki ‘legs’
Most languages that allow reduplication allow both full and partial reduplication. Out of the sample of 368 languages in the World Atlas of Language Structures,7 only 9.5% allowed full reduplication only. No language in this sample allowed only partial reduplication.
In the case of full reduplication, there is no variation in what the reduplicant looks like, as the reduplicant cannot be anything other than equal to the base. But, for partial reduplication, there are various types, depending on which part in particular is copied.
One way to think of partial reduplication is a tension between the need to preserve material from the base and the need for the reduplicant to conform to a particular phonological shape.
The shape demanded by the reduplicant can be a set of segments (C, CV, CVCV), a set of syllables, or a set of morae. Here are some examples of various types of partial reduplication found in Ilocano, where the reduplicant copies a different amount of material from the base in each case.
Ilocano (Austronesian, Northern Philippines)8
V- kumrad → kumra<a>d 'creak'
C- lalaki → la<l>laki 'boys'
CV- nuang → nu-nuang 'water buffaloes'
CVC- bato → bat-bato 'stones'
CVCV- tapiken → tapi-tapikean 'pat repeatedly'
Most often in partial reduplication, the shape of the reduplicant is defined in terms of morae or syllables rather than a specific number of consonants and vowels. As a result, the shape of the reduplicant can vary from word to word.
For example, Mokilese progressive reduplication requires a bimoraic reduplicant. When the initial syllable in the base is itself bimoraic, as with the word soorↄk 'tear', this is copied over exactly, as in soo-soorↄk 'tearing'. When the first syllable is monomoraic, a consonant from the onset of the next syllable is copied as well. So the base pↄdok has the reduplicant pↄd, because this constitutes a bimoraic syllable in isolation, despite the fact that pↄdok is syllabified pↄ.dok. Finally, when there is no following consonant available to create a bimoraic reduplicant, the base vowel is lengthened: diar 'find' → dii-diar 'finding'.
pↄdok 'plant' → pↄd-pↄdok 'planting'
soorↄk 'throw' → soo-soorↄk 'throwing'
diar 'find' → dii-ar 'finding'
Morphological conditions on reduplication
The material selected for copying by the reduplicant can also be defined morphologically.
For example, in languages with morphologically complex stems, reduplication can target various levels in that stem. The following examples are drawn from the Bantu family, and the difference in the size of the reduplicant is easily noticed here, as the roots involved are etymologically related.
In Ciyao, the entire stem of the base, including its suffixes, is copied to form the reduplicant.
Ciyao (Bantu P.21)10
telec-el-a 'cook-APPL-FV'11 → telecela-telec-el-a 'cook for someone frequently'
dim-ile 'cultivate-PERF' → dimile-dim-ile 'cultivated many times'
In the equivalent construction in Ndebele, by contrast, the applicative suffix is reduplicated, while the final vowel or subjunctive suffix is not.
Ndebele (Bantu S.44)12
lim-el-a 'cultivate-APPL-FV' → lime+lim-el-a 'cultivate a little, here and there'
lim-e 'cultivate-SUBJ' → lima+lime13 'cultivate a little, here and there (subjunctive)'
And in Kinyarwanda, only the root is reduplicated – none of the base's suffixes make it into the reduplicant.
The existence of phonological well-formedness conditions on the reduplicant that give rise to partial reduplication can also give rise to other complexities. One such complexity is the fact that phonological conditions on the reduplicant may cause material to be included in the reduplicant which is not present in the base.
An example of such a well-formedness requirement can be found in Kinande, where there is a reduplication construction which normally takes as its base only the root and not the noun class prefix. However, if the root is less than two syllables long, the noun class prefix gets copied as well:
Kinande (Bantu, J.42)16
a. kʊ-gʊlʊ ‘leg’ → kʊ-gʊlʊ-gʊlʊ ‘a real leg’
b. ri-bwɛ ‘snake’ → ri-bwɛ-ribwɛ ‘a real snake’
Similarly, in Emerillon, reduplication indicating repeated action normally takes as its base the root alone (e.g. ʤika, eta, Ɂal, wag below), but will include prefixal material (o-, lo-) as necessary to fill out two syllables.
o-ʤika-ŋ '3-kill-PL' → o-ʤika-ʤika-ŋ
o-eta '3-cut' → o-eta-eta
o-Ɂal-oŋ '3-fall-PL' → oɁa-o-Ɂal-oŋ
a-lo-wag '1sg-CAUS.COM-go' → a-lowa-lo-wag
Simple and complex reduplication
Reduplication can be simple or complex. Everything we've discussed thus far has been simple alliteration: simple reduplication involves repeating some amount of material from the base without alteration. Complex reduplication involves altering the base material before repeating it. Just as we've seen examples of full and partial simple reduplication, there is also both full and partial complex reduplication.
Full complex reduplication constructions are often called "echo constructions", "echo reduplication", or "alliterative repetition". In English, schm-reduplication (Hollywood, Schmollywood!) is an example of full complex reduplication. Another example can be found in Persian:
Persian (Indo-European, Iranian)18
bâlâ ‘above’ → bâlâ-mala ‘somewhere above’
mive ‘fruit’ → mive-pive ‘fruit and so on’
As for partial complex reduplication, one such example can be found in Thao Ca-reduplication, where the first consonant of the base is copied, and the invariant vowel a is added to it.
Thao (Austronesian, Paiwanic)19
finshiq 'to sow' → fa-finshiq ‘seed for planting’
cput 'to filter' → ca-cput 'sieve'
kishkish 'to shave, cut' → ka-kishkish 'razor'
Another example of partial complex reduplication is found in Nakanai, where the vowel in the CV-reduplicant is changed to the vowel preceding the reduplicant in the base.
Nakanai (Austronesian, Oceanic)20
haro ‘day’ (sg) → ha<ra>ro ‘days’ (pl)
velo ‘bubbling’ → ve<le>lo ‘bubbling forth’
hilo ‘see’ → hi<li>lo ‘seeing’ (prog)
baharu ‘widow’ → ba<ha>raru ‘widows’ (pl)
For a third example of partial complex reduplication, consider the Ilocano CVCVN-reduplication construction, where N stands for a nasal which is added to the CVCV portion of the reduplicant which comes from the base. This nasal then assimilates to the place of articulation of the following consonant.
Ilocano (Austronesian, Northern Philippines)21
CVCVN- rupa → rupan-rupa 'face to face'
pateg → patem-pateg 'mutual caring'
Now that we've discussed the two main ways of classifying reduplication, namely full vs. partial and simple vs. complex, let's move on to some other phenomena related to reduplication.
Reduction of the reduplicant
In partial reduplication, it is common for the reduplicant to be simpler than the base: contrasts possible in the base are neutralized in the reduplicant. The contrasts neutralized in the reduplicant may be between different syllable shapes, vowel lengths, or different segments. In this way, a language may:
- allow closed syllables in general, but restrict the reduplicant to open syllables.
- allow length contrasts in general, but restrict the reduplicant to short vowels.
- contrast five vowels in general, but restrict the reduplicant to only one of two vowels.
For example, the reduplicant in the Sanskrit intensive reduplication construction simplifies onset clusters in the base.
a. krand ‘cry out’ → kan-i-krand
b. dhvans 'sound' → dhan-i-dhvans-
Despite being well-known in partial reduplication, these reduction effects are generally not found in full reduplication.
Automatic reduplication occurs when another affix triggers reduplication of the base. For example, the Tagalog prefix ka-, which creates adjectives, reduplicates the first two syllables of the word it is affixed to.
Tagalog (Austronesian, Meso-Philippine)23
wilih 'interested' → ka-wilih-wilih 'interesting'
panabik 'excitement' → ka-pana-panabik 'exciting'
The reduplicant is usually attached directly to the base. In some cases, however, a linker intervenes, creating discontinuous reduplication. In the Manila Bay Creoles, there is discontinuous reduplication where the linker -ng- is inserted between the reduplicant and the base:
Manila Bay Creoles (Creole, Spanish-Lexified)24
buníta 'beautiful' → bunita-ng-buníta 'very beautiful'
Direction of reduplication
The reduplicant most often appears to the left of the base, but in some cases also appears in the middle of the base or after the base.25 Partial reduplication tends to occur on the opposite edge from the side at which affixation tend to occur in a language.26 As a result, suffixing languages tend to have prefixing reduplication; prefixing languages tend to have suffixing reduplication.
Functions of reduplication
Reduplication is associated with many different grammatical functions. Cross-linguistically, the most common functions of reduplication are to express meanings like: pluralization, repetition, frequency, and emphasis.
These are all meanings that echo the form of reduplication. This match between form and function is called iconicity. In creole languages, reduplication is most often used in an iconic fashion.27 Sign languages also use reduplication,28 often with the iconic meanings of plurality (for nouns) or, for verbs, meanings such as durative, iterative, or habitual aspect.29
Pangasinan (Austronesian) plural reduplication30
manók ‘chicken’ → manó-manók ‘chickens’
Luiseño (Uto-Aztecan) iterative reduplication31
lawi ‘to make a hole’ → lawa-láwi ‘to make many holes, more than two’
Choctaw (Muskogean) durative reduplication32
tonoli ‘to roll’ → to<no>noli ‘to roll back and forth’
Hunzib (Nakh-Daghestanian) emphatic reduplication33
bat’iyab ‘different’ → bat’-bat’iyab ‘very different’
muǧáƛ ‘after’ → mu-muǧáƛ 'much later'
But reduplication serves many non-iconic functions in the world's language as well. Reduplication functions in ways associated with derivational morphology. For example, reduplication can change lexical category In the Oceanic language Banoni, partial reduplication is used to convert verbs into nouns, while in another Oceanic language, Ulithian, reduplication turns nouns into verbs.
resi ‘grate coconut’ → re-resi ‘coconut grater’
sogu ‘to husk coconut’ → so-sogu ‘coconut-husking stick’
sifu ‘grass skirt’ → sif-sifu ‘wear a grass skirt’
yaŋi ‘wind’ → yaŋi-yaŋi ‘blow’.
Reduplication has been known to mark valency changing operations on verbs, such as in Nadrogā,36 where it forms intransitives out of patient-oriented verbs.
Nadrogā (Austronesian, Oceanic)37
vuli 'to be turned over' → vuli-vuli ‘turn over’
Reduplication can also serve perform inflectional functions unrelated to plurality. For example, Tarok uses reduplication to express agreement with a third-person singular possessor.
Tarok (Banue-Congo, Platoid)38
a-fini ‘yarn’ → a-_fini-fini ‘his/her yarn’
a-dànkali ‘potato’ → a-dànkali-dànkali ~ a-dànkali-kali 'his/her potato'
Not all morphological categories have been known to be expressed by reduplication. There are few examples, if any, of reduplication being used to mark applicatives, negation, or case marking.
Triplication is similar to reduplication, except that in triplication the reduplicant occurs twice. Triplication is not particularly common. In at least some of the rare cases where triplication does occur, it is used with distinct meaning from reduplication. For example, in Mokilese, reduplication is are used to express that an action is repeated many times, and triplication to express that an action is repeated even more times.
Mokilese (Austronesian, Eastern Malayo-Polynesian)39
roar ‘give a shudder’ → roar-roar ‘be shuddering’ → roar-roar-roar ‘continue to shudder’
Triplication is also used in certain Sino-Tibetan languages. In the Mazi dialect of Stau (Sino-Tibetan, Rgyalrongic), verbs reduplicate to agree with plural subjects. So for a verb ndzu ‘sit’, the third-person plural is ndzə-ndzu ‘they sit’. Note that the vowel u in the base ndzu gets replaced with ə in the reduplicant portion ndzə-.
This illustrates the phenomenon of the reduction of the reduplicant: distinctions allowed in the base are neutralized in the reduplicant. So, while many vowels are possible in the base, only ə (and i) are found in the reduplicant. Another example is zdær ‘drip’, which becomes zdə-zdær ‘they drip’ in the third-person plural. Here we see another common feature of reduplicants: in addition to neutralizing distinctions found in the base, reduplicants often only permit a simpler structure than is allowed in the base: in zdə-zdær, the reduplicant zdə- omits the final consonant r found in the base zdær.
Stau has another form used for plural subjects, which features triplication: ndzə-ndzə-ndzu ‘they sit’. Here we see the base ndzu preceded by two copies of the reduplicant ndzə-. The second reduplicant is subject to the same restrictions as the first: it must be an open syllable (i.e. no final consonant) and it must have the vowel ə.
Like reduplication, triplication in Mazi Stau is used to indicate verb agreement with plural subject. The triplicated form is used to emphasize the fact that multiple subjects are involved, or if the number of subjects has increased, you may opt to use the reduplicated or triplicated form. The triplicated form in particular is used when the subject refers to more than 10 individuals, especially when the large number of individuals has some importance for the discourse.
Triplication in general occurs in a few other Sino-Tibetan languages apart from Stau,30 including Tibetan40 and Taiwanese.27 The Tibetan examples are all onomatopoeic words, such as the names of birds. In Taiwanese, triplication serves to intensify adjectives.
Triplication seems much more closely associated with the concept of ‘more’41 than reduplication does, whether it expresses the intensification of an adjective (as in Taiwanese) or an increase in the number of subjects a verb agrees with (as in Stau).
- Bakker, P. and M. Parkvall (2005). Reduplication in pidgins and creoles. Studies on reduplication, B. Hurch and V. Mattes, eds. Berlin: Mouton de Gruyter: 511–531.
- Blevins, Juliette (1996). Mokilese reduplication. Linguistic Inquiry 27: 523–530.
- Blust, R. (1995). Notes on Berawan consonant gemination. Oceanic Linguistics 34(1): 123–38.
- Chang, L. 1998. Thao reduplication. Oceanic Linguistics 37(2): 277–297.
- Cohn, Abigail (1989). Stress in Indonesian and bracketing paradoxes. Natural Language and Linguistic Theory 7: 167-216.
- Downing, Laura (2000). Morphological and prosodic constraints on Kinande verbal reduplication. Phonology 17: 1–38.
- French, K. (1988). Insights into Tagalog. Reduplication, Infixation, and Stress from Nonlinear Phonology. SIL.
- Gates (2017). Verbal triplication morphology in Stau (Mazi dialect). Transactions of the Philological Society 15(1): 14–26.
- Geraghty, Paul (2002). Nadrogā. In The Oceanic Languages, J. Lynch, M. Ross, and T. Crowley, eds. Curzon Press: 833–847.
- Ghaniabadi, S., J. Ghomeshi and N. Sadat-Tehrani (2006). Reduplication in Persian. A morphological doubling approach. In Proceedings of the 2006 Canadian Linguistics Association Annual Con- ference, C. Gurski and M. Radisic, eds.
- Grant, A. (2003). Reduplication in Mindanao Chabacano. In Twice as Meaningful. Reduplication in Pidgins, Creoles and Other Contact Languages, S. Kouwenberg, ed. Battlebridge: 203–10.
- Harrison, Sheldon P. (1973). Reduplication in Micronesian languages. Oceanic Linguistics 12: 407–454.
- Harrison, Sheldon P. (1976). Mokilese Reference grammar. The University Press of Hawaii.
- Haspelmath, Martin (2013). Functions of reduplication. In The atlas of pidgin and creole language structures (Susanne Maria Michaelis, Philippe Maurer, Martin Haspelmath, Magnus Huber, eds.). Oxford University Press.
- Hyman, Larry M. (2008). The natural history of reduplication in Bantu. UC Berkeley Phonology Lab Annual Report 2008: 153–187.
- Hyman, Larry M., Sharon Inkelas, and Galen Sibanda (2009). Morphosyntactic correspondence in Bantu reduplication. The Nature of the Word: Essays in Honor of Paul Kiparsky, K. Hanson and S. Inkelas, eds. MIT Press.
- Inkelas, Sharon (2012). Reduplication. In The handbook of exponence, J. Trommer, ed. Oxford University Press.
- Johnston, R. 1980. Nakanai of New Britain. The Grammar of an Oceanic Language. The Australian National University.
- Kimball, Geoffrey D. 1988. Koasati Reduplication. In In Honor of Mary Haas: from the Haas Festival Conference on Native American Linguistics, William Shiplay, ed. Mouton de Gruyter: 431–442.
- Kimenyi, Alexandre. 2002. A tonal grammar of Kinyarwanda: an autosegmental and metrical analysis. (Studies in linguistics and semiotics, 9.) Edwin Mellen Press.¯
- Kroeber, A. L. and George William Grace (1960). The Sparkman Grammar of Luiseño. (University of California Publications in Linguistics, 16.) University of California Press.
- Lin, Paul Y. (1982). Semiotic perspectives on Chinese: A picturesque language. In Semiotics 1980: Proceedings of the Fifth Annual Meeting of the Semiotic Society of America, Michael Herzfeld & Margot D.Lenhart, eds. Plenum: 281–296.
- Lincoln, Peter (1976). Describing Banoni, an Austronesian language of Bougainville. PhD dissertation, University of Hawaii.
- Lynch, John (2002). Ulithian. In The Oceanic Languages, J. Lynch, M. Ross, and T. Crowley, eds. Curzon Press: 792–803.
- Lynch, John and Malcolm Ross (2002). Banoni. In The Oceanic Languages, J. Lynch, M. Ross, and T. Crowley, eds. Curzon Press: 440–455.
- Marantz, Alec (1982). Re reduplication. Linguistic Inquiry 13: 483–485.
- McCarthy, J. & E. Broselow (1983). A theory of internal reduplication. Linguistic Review 3(1): 25–88.
- Mutaka, Ngessimo, and Larry M. Hyman (1990). Syllable and morpheme integrity in Kinande reduplication. Phonology 7: 73–120.
- Ngunga, Armindo (2001). Phonology and morphology of the Ciyao verb. CSLI Publications.
- Niepokuj, Mary Katherine (1997). The development of verbal reduplication in Indo-European. Institute for the Study of Man.
- Perniss, Pamela, Roland Pfau, and Markus Steinbach (2007). Can’t you see the difference? Sources of variation in sign language structure. In Visible Variation: Cross-Linguistic Studies on Sign Language Structure (Pamela Perniss, Roland Pfau, and Markus Steinbach, eds.). Mouton de Gruyter: 1–34
- Robinson, J. O. (1976). His and hers morphology: the strange case of Tarok possessives. Studies in African Linguistics, Supplement 6, L. Hyman, L. Jacobson and R. Schuh, eds. Department of Linguistics, UCLA. 201–209.
- Rose, Francoise (2005). Reduplication in Tupi-Guarani languages. In Studies on Reduplication, B. Hurch, ed. Mouton de Gruyter: 351–368.
- Rubino, Carl (2000). Ilocano Dictionary and Grammar. Ilocano-English, English-Ilocano. University of Hawai‘i Press.
- Rubino, Carl (2001). Pangasinan. In Encyclopedia of the World's Languages: Past and Present, Jane Garry and Carl Rubino, eds. H.W. Wilson Press: 539–542.
- Rubino, Carl (2011). Reduplication: Form, function and distribution. In Studies on Reduplication (Bernard Hurch, ed.). De Gruyter Mouton.
- Rubino, Carl (2013). Reduplication. In The World Atlas of Language Structures Online (Matthew S. Dryer and Martin Haspelmath, eds). Max Planck Institute for Evolutionary Anthropology.
- Sibanda, Galen (2004). Verbal phonology and morphology of Ndebele. PhD dissertation, University of California, Berkeley.
- Sibomana, Leonidas (1980). Grundzüge der Phonologie des Tarok (Yergam). Afrika und Übersee 63: 199-206.
- Sibomana, Leonidas (1981). Tarok II: Das Nominalklassensystem. Afrika und Übersee 64: 25–34.
- Steriade, Donca (1988). Reduplication and syllable transfer in Sanskrit and elsewhere. Phonology 5: 73–155.
- Uray, Géza (1954). Duplication, germination and triplication in Tibetan. Acta Orientalia Academiae Scientiarum Hungaricae 4.1-3: 177-256.
- van den Berg, Helma (1995). A Grammar of Hunzib (Lincom Studies in Caucasian Linguistics, 1). Lincom Europa.
- Velupillai, Viveka (2012). An Introduction to Linguistic Typology. John Benjamins.
Much of the data in this article comes from Velupillai (2012: 100–104) and Inkelas (2012). If you'd like a more thorough introduction to the theoretical issues raised by reduplication, I recommend reading the latter work especially. ↩
This generalization extends only to the spoken languages of Western Europe. ↩
Rubino (2011). ↩
Haspelmath (2013). ↩
Roberts (1987: 252–254, 1991: 130–131), as cited in Inkelas (2012). ↩
Cohn (1989:185), as cited in Inkelas (2012). ↩
Rubino (2013). ↩
Rubino (2000: xvii), as cited in Velupillai (2012: 102–103). ↩
Blevins (1996:523), citing Harrison (1973, 1976), as cited in Inkelas (2012). ↩
Ngunga (2001), as cited in Inkelas (2012). ↩
FV = final vowel. ↩
Sibanda (2004), as cited in Inkelas (2012). ↩
The final vowel a of the reduplicant in the form lima-lime is added to fulfil a condition that the reduplicant must be disyllabic (Inkelas 2012: 5). ↩
Kimenyi (2002), as cited in Inkelas (2012). ↩
As in Ndebele, the final vowel aa of the reduplicant in Kinyarwanda is inserted to fulfill the disyllabic reduplicant condition. ↩
Mutaka & Hyman (1990: 77–80), Downing (2000), Hyman (2008), as cited in Inkelas (2012). ↩
Rose (2005: 353–359), as cited in Inkelas (2012). ↩
Ghaniabadi et al. (2006: 3), as cited in Velupillai (2012: 102). ↩
Chang (1998: 282), as cited in Velupillai (2012: 102). ↩
McCarthy & Broselow (1983: 74), citing Johnston (1980: 149), as cited in Velupillai (2012: 103) ↩
Rubino (2000: xvii), cited in Velupillai (2012: 102–103). ↩
Steriade (1988: 108), as cited in Inkelas (2012). ↩
French (1988: 50), as cited in Velupillai (2012: 103). ↩
Grant (2003: 205), as cited in Velupillai (2012: 103). ↩
Rubino (2013). ↩
Hyman (2008), Hyman et al. (2009). ↩
Perniss et al. (2007), as cited in Velupillai (2012). ↩
For an action to be durative signifies that it takes place over a long period of time; iterative means that the action happens repeatedly; habitual means that it takes place habitually. ↩
Kroeber and Grace (1960). ↩
Kimball (1988: 440). ↩
van den Berg (1995: 34), as cited in Rubino (2013). ↩
Lynch (2002:799), as cited in Inkelas (2012). ↩
Lynch & Ross (2002:442), Lincoln (1976:164), as cited in Inkelas (2012). ↩
in Nadrogā is pronounced with the velar nasal [ŋ]. ↩
Geraghty (2002: 841), as cited in Inkelas (2012). ↩
Niepokuj (1997:23), citing Robinson (1976), Sibomana (1980, 1981), as cited in Inkelas (2012). ↩
Harrison (1973), as cited in Rubino (2013). ↩
Uray (1954). ↩
Blust (2001: 335). ↩