Thulung Rai, an endangered Tibeto-Burman language of Eastern Nepal, has complex verbal morphology, with verb endings encoding agent and patient person and number in transitive scenarios. In addition to this, a large number of verbs alternate between several stems, and the stem selection criteria are initially elusive. Inspired by work by Boyd Michailovsky, who proposes morphophonological accounts for the verb stem alternation in related Dumi Rai, I propose an analysis of the Thulung verbal system and its verb stem alternation.
whose comments on an earlier draft vastly enhanced the analysis presented here, as well as the clarity of the overall paper. I also thank the two anonymous reviewers and the journal's editors for very constructive comments. I am of course solely responsible for any mistakes in the data and analysis presented here. I am also very grateful to LACITO-CNRS for funding a research trip to Nepal where a good part of the material for this analysis was collected.
1 The following abbreviations are used: DE=dual exclusive; Di=dual inclusive; Du=dual; ERG=ergative; Hs=hearsay iNF=infinitive; MM=middle marker; rrPS'i`=non- past; PE=plural exclusive; PI=plural inclusive; PL=plural; PST-past; sc-singular; x> Y = X agent acting on Y patient.
2 A default 3SG patient is given for all the transitive paradigms in this paper, this being by far the most common form in narrative contexts.
' Excepting allomorphy in the Isg and 3PL past verb endings, which have an easily identifiable distribution: in vowel-final environments, ISG>3SG.PST is -uto (vs. post-consonantal -to), IsG.PST is -gro (vs. post-consonantal -goro), 3PL.PST and 3pl>3sG.PST are -mri(vs. post-consonantal -min).
4 Note that the change is unexpected, given that in pre-consonantal environments, *t is the reflex of an initial but rather of a final.
5 This is true throughout the language, except when the medial consonant is an s. This is presumably a matter of syllabification: the s clusters with the following consonant and is a syllable onset. 6 But note that Thulung does not use metathesis as part of its grammar synchronically.
7Allen (1975:61) also organizes his data into stem classes, calling them "types of stem", but also referring to a stem type as forming a "class". He provides a grid (Table IV on p. 62) where he shows what happens to each Cf when it occurs with a specific "ending class" (somewhat abstracted endings outlined in his Table I on p.46). This analysis of verb stem alternation in terms of the Cf of the verb is adopted in most grammars of Kiranti languages. In some cases, the observed alternation is straightforwardly phonological, so that the two types of stems (pre- consonantal and pre-vocalic) are often related through lenition or reduction. This is, for instance, the case in Camling (Ebert 1997), Athpare (Ebert 1997), Yamphu (Rutgers 1998), and Bantawa (Doomenbal 2009). Other languages present more complex systems of alternation, where phonological criteria alone cannot explain the synchronic alternation patterns. For example, an ending of identical form can be found with different stems of the same verb. This is the case in Dumi (van Driem 1993), Wambule (Opgenort 2004), Jero (Opgenort 2005), and Koyi (Lahaussois 2009). For such languages, one way to present the data is by means of paradigms showing co-occurrence patterns of stems with person/number/tense endings. In some cases, the data cannot be accounted for in terms of a set of rules or regularities, because the factors conditioning the attested alternations are no longer recoverable. In other cases, morphophonological conditions can be proposed that elucidate the observed phenomena. This is the case of Thulung, as discussed presently.
8 We saw above that there is no Cf tin Thulung (we get (instead). These verb stem class labels are the same as used by Allen (1975:62). 9 We get full stem CVd -- CVr- in pre-vocalic environments, and full stem CVt- pre-consonantally.
10 See Ohala and Ohala (1993:234-5) for acoustic explanations of why back nasals are less consonantal than front nasals. This has to do with the shortness of the oral resonating cavity for velar nasals in comparison with labial and alveolar nasals, and the result is that back nasals are often less common, sometimes even alternating with nasalized vowels. I do not find nasalized vowels in the places where an expected velar nasal does not appear, for example in k stem class weakened stems, but they may have been nasalized at some earlier time.
" I use these labels in what is admittedly a somewhat clumsy fashion to distinguish between an intransitive subject ('subject') and a transitive subject (` agent' ). 12 This surface similarity led me to conclude in my earlier work (Lahaussois 2002:155) that there was no phonological basis for the stem alternations, as the same endings were found with different stems.
13 Note how this is distinct from a non-alternating nasal final verb paradigm such as that in Table 8.
'4 While this might suggest that the basic distribution is simply one of a weakened stem for the non-past and a full stem for the past, such a hypothesis makes it quite challenging to account for the presence of the full stem with specific c person/number combinations in the non-past. If one were to imagine that the "past stem" had the form it does as a result of the presence of a past marker that would suggest that the default stem was the "non-past stem". Considering that the relationship between full and weakened is one of lenition, it seems more efficient to claim that the default stem is the full stem, and that phonological processes account for the situations where we find something other than the full stem.
15 As shown in Section 4.1.3, these endings are vowel-initial, even if their surface form sometimes contains epenthetic material (-p for lsG, -4 for IPI and 3sG) and therefore looks consonant-initial.
16 Descriptions of some Kiranti languages mention a stem distribution whereby a "pause" or "word boundary" is preceded by the pre-consonantal stem, this is the case of Chamling (Ebert 1997b:14), Athpare (Ebert 1997a:20) and Yamphu (Rutgers 1998:103). To quote Ebert's (1997b:14) analysis of Chamling, "[t]hc full stem occurs only before vowels. Before a consonantal suffix or a pause most stems are reduced or undergo a morphophonological change." In these languages, ante- vocalic stems cannot be realized without a subsequent vowel due to resyllabification. It is nonetheless interesting that in these languages, as in Thulung, it is the weakened stem that occurs in an environment where there is no suffix to condition the weakening.
17 Curiously, some other cross-linguistically middle marked semantic situation types, such as grooming and body care, are not marked with -s in Thulung but rather with reflexive marker -si. Yet note that according to Kemmer (1993:55), "[grooming actions] are of central importance as they represent a situation type that is very frequently, if not universally, middle-marked in languages with middle marking.'
ALLEN N.J. (1975). Sketch of Thulung Grammar, with Three Texts and a Glossary. Ithaca: China-Japan Program.
BORCHERS Dörte (2008). AGrammar of Sunwar: Descriptive Grammar,Paradigms, Textsand Glossary. Brill's Tibetan Studies Library. Languages of the Greater Himalayan Region, 5,7. Leiden: Brill.
DEWSA RAI, Agam Sing (1944). Asalachhi Shiksa, ThulungRai Bhasa. Darjeeling.
DRJEM van George (1993). A GrammarofDumi. Berlin: Mouton de Gruyter.
DOORNENBAL Marius (2009). A Grammar ofBantawa:Grammar,ParadigmTables, Glossary and Texts of aRai LanguageofEasternNepal. Utrecht: LOT
EBERT Karen (1994). The Structure of theKiranti Languages. Arbeiten des Seminar fur Allgemeine Sprachwissenschaft Nr. 13. Zürich: Universitat Zürich Seminar fur Allgemeine Sprachwissenschaft.
EBERT Karen (1997a). Athpare. Munich: Lincom Europa.
EBERT Karen (1997b). Camling (Chamling). Munich: Lincom Europa.
EBERT Karen (2003). Kiranti languages: An overview. In: THURGOOD G. and R. LAPOLLA (eds.). The Sino-TibetanLanguages. London and New York: Routledge. Pp. 505-517.
KEMMER Suzanne (1993). TheMiddleVoice. (Typological Studies in Language 23). Amsterdam and Philadelphia: John Benjamins.
KEMMER Suzanne (1994). Middle voice, transitivity and the elaboration of events. In: FOX B. and P. HOPPER (eds.). Voice:FormandFunction. Amsterdam and Philadelphia: John Benjamins. Pp. 179-230.
LAHAUSSOIS Aimee (2002). Aspects of the Grammar of ThulungRai:anEndangered Himalayan Language. PhD dissertation: University of California, Berkeley.
LAHAUSSOIS Aimee (2003). Thulung Rai. Himalayan Linguistics Archive I. Pp. 1-25. http://www.linguistics.ucsb.edu/Himalayan Linguistics/grammars/index.html.
LAHAUSSOIS Aimee (2009). Koyi Rai: an Initial Grammatical Sketch. Himalayan Linguistics Archive 4. Pp. 1-33. http://www.linguistics. ucsb.edu/HimalayanLinguistics/grammars/index.html
MICHAILOVSKY Boyd (1975). Notes on the Kiranti verb (East Nepal). Linguisticsof the Tibeto-BurmanArea, vol. 2, no. 2, pp. 138-218.
MICHAILOVSKY Boyd (1994). Manner vs Place of Articulation in the Kiranti Initial Stops. In: KITAMURA H. et al. (eds.). Current IssuesinSino-TibetanLinguistics. Osaka: National Museum of Ethnology. Pp. 766-772, 781.
MICHAILOVSKY Boyd (2009). Preliminaries to the comparative study of the Kiranti subgroup of Tibeto-Burman. Proceedings of the International Symposium on Sino-Tibetan Comparative Studies in the 21st Century, June 24-25, 2010. Academia Sinica, Taipei, Taiwan. Pp. 145-70.
MICHAILOVSKY Boyd. (Ms). The Dumi verb revisited: lexical bases and stem formation.
OHALA J. J. and OHALA M. (1993). The phonetics of nasal phonology: theorems and data. In: HUFFMAN M.K. and KRAKOW R.A. (eds.). Nasals, Nasalization, and the Velum. (Phonetics and Phonology Series, Vol. 5.) San Diego, CA: Academic Press. Pp. 225-249.
OPGENORT Jean-Robert(2004). A Grammar ofWambule.Grammar,Lexicon,Texts and Cultural Survey of aKirantiTribe ofEastern Nepal. Brill's Tibetan Studies Library. Languages of the Greater Himalayan Region, 2. Leiden: Brill.
OPGENORT Jean-Robert (2005). A GrammarofJero. WithaHistoricalComparative Study of theKirantiLanguages. Brill's Tibetan Studies Library. Languages of the Greater Himalayan Region, 3. Leiden: Brill.
RUTGERS Roland (1998). Yamphu: Grammar, Texts andLexicon. Leiden: Research School CNWS.
TOLSMA Gerard (2006). A GrammarofKulung. Brill's Tibetan Studies Library. Languages of the Greater Himalayan Region, 4. Leiden: Brill.