Associated motion in Sino-Tibetan/Trans-Himalayan

Associated motion in Sino-Tibetan/Trans-Himalayan* Guillaume Jacques (CNRS-CRLAO) Aimée Lahaussois (CNRS-HTL) Zhang Shuya (INALCO-CRLAO) December 1, 2017 1 Introduction This article aims to examine associated motion (henceforth AM) in the Sino-Tibetan languages and to determine the distribution of the phenomenon across a number of language groups of the family. We also aim to look into the contribution that the study of the phenomenon in Sino-Tibetan languages can bring to a wider typological discussion of this linguistic category, through the consideration of additional parameters and questions that have not yet been taken up in the existing literature. The language groups considered in this study are Gyalrongic, Kiranti, Sinitic, as well as a few additional languages where AM has been found. For each of the language groups, we examine the main characteristics of AM marking, highlighting what is distinct from the presentation of the phenomenon in the other ST groups in the sample. We ind that AM is present in a number of groups within the ST family, but that the conigurations of AM are quite diferent from one group to the next. The comparison grid that we apply in looking at the various language groups is based on previously identiied parameters of AM which have been described in the literature for diferent linguistic areas (in particular Koch 1984, Wilkins 1991 and Guillaume 2016), namely temporal relation, deixis and argument of motion, in addition to a few extra parameters which are particularly relevant for the Sino-Tibetan languages (Vertical dimension, Mono- vs pluriactionality and Volitionality) The vertical dimension is a parameter which is particularly relevant to the Kiranti subgroup, with most of the languages in the group have AM markers which encode ’upwards’, ‘horizontal’ and ’downwards’ orientation into motion verbs. Another parameter we have found to be relevant in the Sino-Tibetan languages is whether the verb+AM marker unit is mono- or pluri-actional, namely whether the motion event can be separated out from the main action or not. A number of tests are proposed to determine the inseparability of motion event and verbal action in diferent languages in § 2.4. Volitionality is considered in § 2.5. In this survey, we consider all markers for which examples show AM, even if the primary function of the marker is to express some other feature, most commonly aspect or orientation. We do not, however, take into account markers which do not encode any AM – even though these may in many cases come from source verbs which indicate motion –, or those for which the examples are questionable as to the interpretation that should be given to the marker in question. In working through the grammars of languages in our sample, we have identiied a number of issues in the presentation of material making it diicult to determine whether or not AM is present in a language, and we thus * The Japhug, Situ, Khaling and Thulung data are from the authors’ ieldwork. The Japhug and Khaling examples are taken from corpora that are progressively being made available on the Pangloss archive (Michailovsky et al. 2014, http://lacito.vjf.cnrs.fr/pangloss/corpus/list_rsc.php?lg=Japhug and http://lacito.vjf.cnrs. fr/pangloss/corpus/list_rsc.php?lg=Khaling). 1 provide a list of suggestions for ieldworkers in order to better capture the phenomenon during future data collection. 2 Gyalrongic 2.1 General overview Associated motion preixes are reported in Japhug (Jacques 2013), Zbu (Gong 2018), Tshobdun (Sun 2014) and Situ (Zhang 2016: 200-204, Prins 2016: 497-500), but not found in other Gyalrongic languages (for example in Khroskyabs, cf Lai 2017). The main characteristics which set apart AM in Gyalrongic languages within the Sino-Tibetan context are the following: 1. The languages have devoted AM preixes. 2. AM preixes and orientation preixes co-occur, illing distinct templatic slots. 3. AM marking is associated with echo phenomena, with multiple marking coding a single motion event. 4. Verbs with AM marking are mono-actional (for Japhug). These characteristics will be discussed in detail after an overview presenting how the main parameters of temporal relation, deixis and argument of motion manifest for Gyalrongic languages. 2.1.1 Temporal Relation AM preixes in Japhug and other Gyalrongic languages refer to a motion event occurring before the action of the main verb, resulting in a prior temporal relation with respect to the main verb, as in (1) and (2). ɣɯ-pjɯ-nɯ-tsʰi-nɯ (1) tɕe tɯ-ci lnk indef.poss-water come&do-ipfv-auto-drink-pl ‘(The wild yaks) come to drink water there.’ (20-RmbroN, 46) (Japhug) ɕ-pjɤ-nɯ-tsʰi. (2) tɕe tɯ-ci lnk indef.poss-water go&do-ifr-auto-drink ‘She went to drink water.’ (140428 mu e guniang, 72) (Japhug) 2.1.2 Deixis All four Gyalrongic languages distinguish cislocative and translocative AM (see examples 1, 2 and 3). They mark the category through two preixes each (see Table 1 from Zhang 2016: 200, Tshobdun data from Sun 2012, Zbu from Gong 2018), grammaticalized from the motion verbs ‘go’ and ‘come’. The fact that motion verbs were grammaticalized as preixes rather than suixes in strict verb-inal languages like Japhug and Situ can be accounted for by assuming that they comes from the irst member of a former serial verb construction (Jacques 2013). AM markers in Gyalrong do not have redundant person and TAM markers, as they do in Kiranti languages. (3) ɐ-kómʔ və-tə-twíʔ 1sg-door come&do-imp-open ‘Come open the door for me!’ (Zbu, Gong 2018) 2 Table 1: Associated motion preixes in Gyalrong languages Japhug Kyom-kyo (Situ) Brag-dbar (Situ) Tshobdun Zbu come ɣi vi βʑē, və wî və̂ cisloc ɣɯvəɟɐovə- go ɕe tʃʰi tɕʰê ʃɐ̂ xwéʔ transloc ɕɯ-, ɕ-, ʑ-,zʃiɕɐʃəɕɯ- Note that in Situ, the cislocative can be used with a prospective aspectual value (Lin 2003, Zhang 2016: 204), whereas in the other Gyalrong languages, it only marks associated motion. 2.1.3 Argument of motion The argument undergoing the motion event is always the subject (S/A), except in the case of causative constructions, when it can be either causer or causee (as in 4 from Japhug and 5 from Tshobdun). (4) tɕe kupa cʰu nɯra atʰi pɕoʁ nɯra, ɯ-pɕi nɯra kɯ kɯreri lnk Chinese loc dem:pl downstream direction dem:pl 3sg-outside dem:pl erg here ŋu. ɣɯ-cʰɯ-sɯ-χtɯ-nɯ come&do-ipfv:downstream-caus-buy-pl be:fact cd ‘People from the Chinese areas, people from outside send people to come here to buy (matsutake and sell them in the areas downstream).’ (20 grWBgrWB 58) (Japhug) (5) tʃone=nəʔ ɐɟiʔ ʃə-te-nɐ́ mɲi-aŋ nɐ-zgɐ̂ t ʃənəʔ show=det 1sg go&do-ipfv-watch-1sg ipfv.pst-be.justiied[II] but mə-ʃə-tə-o-s-nɐmɲi-aŋ neg-go&do-pfv-inv-caus-watch[II]-1sg ‘I deserved to go and watch the show, but s/he did not let me go watch it.’ (Tshobdun, Sun 2012: 478) 2.1.4 Motion verbs and AM preixes In Gyalrongic, there is no constraint on AM preixes occurring on motion verbs with the same deixis. Examples (6) and (7) respectively illustrate the cislocative on the verb ɣi ‘come’ and the translocative on the verb ɕe ‘go’. Such examples are not common enough to allow a clear analysis of the semantic value of the redundant AM in these examples. (6) <jiazhang> ra ju-ɣi-nɯ tɕe <laoshi> ɯ-ɕki, tɯ-ɕki ʑo parents pl ipfv-come-pl lnk teacher 3sg.poss-dat genr.poss-dat emph ɣɯ-ju-ɣi-nɯ ɕti netɕi? come&do-ipfv-come-pl be.affirm:fact sfp ‘The parents come, come to the teachers (us).’ (conversation140501 01, 60) (Japhug) (7) li nɤki iɕqʰa nɯ tɤjlu kɤ-rku ɯ-ŋgɯ zɯ again dem the.aforementioned dem lour nmlz:P-put.in 3sg.poss-inside loc ɕ-pjɤ-ɕe go&do-ifr:down-go ‘He went into the bag of lour.’ (140519 chou xiaoya-zh, 145) (Japhug) 3 The opposite combinations, namely cislocative with ɕe ‘go’ and translocative with ɣi ‘come’, are not grammatical. 2.2 Orientation and AM The irst characteristic of note is that Gyalrongic languages have devoted AM markers. These markers are only used for AM, and occur in a distinct slot within the verbal template. Orientation markers can thus be reliably distinguished from AM markers, clarifying the function of the two markers when they co-occur. Non-orientable verbs (verbs expressing actions other than motion, manipulation, sight or actions with a single direction) select one or two lexicalized orientation preixes. For instance, in Japhug the verb mɯrkɯ ‘steal’ occurs with the orientation ‘up’ (with the orientation preixes tɤ-, ta-, tu-, to-). When non-orientable verbs occurs with AM, the verb normally keeps the lexicalized orientation preix, as in 8, where mɯrkɯ ‘steal’ is used with the tu- ‘up’ preix; the orientation preix is thus irrelevant to the motion event. (8) kɯ-nŋo nɯ qʰe ci ci ɕ-tu-mɯrki kɯ-fse nmlz:S/A-be.defeated dem lnk one one go&do-ipfv-steal[III] nmlz:S/A-be.like ma nɯ ma mɯ-ɲɯ-ɤʁe. apart.from dem apart.from neg-sens-have.to.eat ‘The (lion) which is defeated steals a little out of it, but apart from that has nothing to eat.’ (20-sWNgi, 65) (Japhug) 2.3 Echo phenomena Previous literature on AM has reported the existence of ‘echo phenomena’ in the use of AM markers (Wilkins 1991: 251, Vuillermet 2012: 681-683, Rose 2015: 128-130, Guillaume 2016: 11), namely that the same motion event can be expressed by more than one AM marker. This phenomenon is common in Japhug narratives. Two subtypes of AM echo can be distinguished. First, in examples such as (9) and (10), a motion verb (ɕe ‘go’ and ɣi ‘come’ respectively) is followed by a verb with an AM preix with the same deixis, though there is a single motion event. (9) tɕʰi ɯ-taʁ to-ɕe tɕe ɕ-tɤ-ru stairs 3sg.poss-on ifr:up-go lnk transloc-up:pfv-look ‘He went up the stairs and looked up.’ (08-kWqhi, 18) (Japhug) (10) kʰa mɯ-pɯ-rɤʑi tɕe tɕe, ftɕar nɯ wuma ʑo βɣɯz pjɤ-rɯŋɯŋɤn house neg-pst.ipfv-stay lnk lnk summer dem really emph badger ifr.ipfv-cause.damage tɕe maka, kɯmtʰoʁ ra kɯnɤ ju-ɣi ɣɯ-ɲɯ-sloʁ pjɤ-ŋu. lnk completely threshold pl also ipfv-come come&do-ipfv-dig.up ifr.ipfv-be ‘He was not home, and that summer badgers were causing a lot of damages, they came and even dug up the threshold of the house.’ (27-spjaNkW, 107) (Japhug) Second, we also ind cases, such as (11), without a motion verb, but with two verbs redundantly preixed with the same AM marker (here ɣɯ-). (11) tɕe a-kʰa ra ɣɯ-ta-rɤroʁrɯz, a-mgo ra lnk 1sg.poss-house pl come&do-pfv:3→3’-tidy 1sg.poss-food pl ŋu ɕi ɣɯ-ta-βzu come&do-pfv:3→3’-make be:fact qu 4 ‘Is it (the neighbour’s wife who took pity on me) and came to tidy my house and make food for me?’ (150827 tianluo, 76) (Japhug) Echo AM is required in serial verb constructions (Jacques 2016: 253-255), as shown in (12), where the verbs stu ‘do like’ and the nɯrtɕa ‘tease’ share the same person (3→3’), TMA (imperfective) and AM (translocative) markers. ɕ-tu-ste tɕe ɕ-ku-nɯrtɕe ra pjɤ-ŋu. (12) kɯra dem:prox:pl go&do-ipfv-do.like[III] lnk go&do-ipfv-tease[III] pl ifr.ipfv-be ‘(The mouse) went and teased (the cat) like that.’ (150902 dashu, 31) (Japhug) 2.4 Pluri- vs monoactionality Another topic of interest in Gyalrong is the issue of the separability of the motion event and the verbal action. Unlike other languages with AM such as Kiranti or Tungusic (Stoynova 2016, Alonso de la Fuente & Jacques to appear), in Japhug the scope of negation, interrogation, conditionals and complement-taking verbs applies to the entire verbal event (motion+main action). In case where the verbal action is outside their scope, an alternative purposive construction with a motion verb (the Motion Verb Construction, henceforth MVC). This is however not common to all Gyalrong languages: in the closely related Situ language, the semantic distinction between AM and the corresponding purposive construction is less clear (§ 2.4.5). To express the meaning of motion prior to an action, associated motion preixes are nearly two times as common as corresponding MVCs in the Japhug corpus. There is however a clear semantic diference between the two constructions, which was briely described in Jacques (2013), but is presented here in more detail. AM and MVC difer from each other in that in the former, the completion of both motion event and verbal action is presupposed (AM is monoactional), whereas in the case of the latter, the two can be separated. This mono- vs. pluractionality contrast is most conspicuous in past perfective forms, and can be observed in four types of constructions: concessives (with negation of the verbal action), interrogatives, conditionals and complement clauses. 2.4.1 Concessive A MVC with the motion verb in perfective form can be followed by a clause negating the purposive action, as in (13). In this example, only the motion is realized, while the action expressed by the verb rtoʁ ‘look’ could not be accomplished. (13) nɤ-kɯ-rtoʁ jɤ-ɣe-a ri, mɯ-nɯ-atɯɣ-tɕi, mɯ-pɯ-ta-mto. 1sg.poss-nmlz:S/A-see pfv-come[II]-1sg lnk neg-pfv-meet-1du neg-pfv-1→2-see ‘I came to see you but I did not see you.’ (Japhug) With the corresponding AM verb form ɣɯ-jɤ-ta-rtoʁ ‘I came to see you’, negating the action of the verb is self-contradictory and nonsensical, and a sentence such as (14) is incorrect. (14) †ɣɯ-jɤ-ta-rtoʁ ri mɯ-pɯ-ta-mto cisloc-pfv-1→2-look lnk neg-pfv-1→2-see Intended meaning: ‘I came to see you but I did not see you.’ (Japhug) Additional minimal pairs of the same type are presented in Jacques (2013: 202-203). 5 2.4.2 Interrogative In interrogative clauses, MVCs are required to express meanings such as ‘What/who have you come/ gone to X’, as in example (15), an example which occurs nine times in the corpus. (15) tɕʰi ɯ-kɯ-pa jɤ-tɯ-ɣe? what 3sg.poss-do pfv-2-come[II] ‘What did you come to do?’ (nine examples in the corpus) (Japhug) The diference between MVC and AM in interrogatives can be illustrated by comparing the minimal pair (15) and (16). Example (15), implies that the addressee has not eaten yet, while (16) with AM can only be used if the food ingestion has already taken place, and requires a diferent translation. (16) tɕʰi ɣɯ-tɤ-tɯ-pa-t what come&do-pfv-2-do-pst:tr ‘What did you do upon coming here?’ (elicited) (Japhug) 2.4.3 Conditional The presuppositional diference between MVC and AM is also perceptible in the protasis of conditional clauses. With MVC in the protasis as in (17), there is no presupposition that the verbal action took place, only the motion event constitutes a condition to the state of afair described in the apodosis. (17) nɤ-wa ɯ-kɯ-rtoʁ mɯ~mɤ-jɤ-tɯ-ɣe nɤ aʑo 1sg.poss-father 3sg.poss-nmlz:S/A-look cond~neg-pfv-2-come[II] lnk 1sg mɯ-pɯ-kɯ-mto-a. neg-pfv-2→1-1sg ‘If you had not come to see your father, you would not have seen me.’ (you saw me, but your father was not here) (Japhug) By contrast, with AM, the verbal action necessarily took place, as in example (18). (18) nɤ-wa mɯ~mɤ-ɣɯ-jɤ-tɯ-rtoʁ nɤ pɯ-sɤzdɯxpa 1sg.poss-father cond~neg-come&do-pfv-2-look lnk pst.ipfv-be.pitiful ‘If you had not come to see your father, he would have felt sorry.’ (but you did saw him, so he does not feel sorry) (Japhug) 2.4.4 Complement clauses In complement clauses, verbs with AM preixes are attested, and complement taking verbs always have scope over both the action of the verb and motion event. In (19), the modal verb cʰa ‘can’ and the double negations (with the speciic meaning ‘cannot help’) have scope over both the motion event and the verbal action – this example is taken from a passage in a story where the king reproaches a small child, who just returned from a mission he himself send him to accomplish, not to have irst come to greet him on his return home; the child says these words to justify why he irst went to see his mother before greeting the king – from this context it is clear that both the motion event (to him mother’s house, explaining the child’s failure to go to see the king) and the action ‘drink milk’ (the reason for that motion event) are equally important to the plot and inseparable. 6 (19) tɯ-nɯ ɯ-kɯ-tsʰi ɲɯ-ɕti-a tɕe, jɤ-azɣɯt-a tɕe, indef.poss-breast 3sg.poss-nmlz:S/A-drink sens-be.affirm-1sg lnk pfv-arrive-1sg lnk nɯ mɯ́ j-cʰa-a tɯ-nɯ ci mɤ-ɕɯ-kɤ-tsʰi indef.poss-breast indef neg-go&do-inf-drink dem neg:sens-can-1sg ‘I am (a toddler) who (still) drinks (his mother’s) milk, when I arrived, I could not help but go to drink milk.’ (Norbzang, 262) (Japhug) In (20), the negated modal verb has also on the action of both the main verb and the motion event – the guards would prevent the main character not only to steal, but also to the where the object to be stolen is found. tɕe, ɕɯ-kɤ-mɯrkɯ ɕti (20) ʁmaʁ χsɯ-tɤkʰar kɯ ɲɯ-ɤz-nɤkʰar-nɯ solider three-rounds erg sens-prog-surround-pl be.affirm:fact lnk go&do-inf-steal mɤ-tɯ-cʰa neg-2-can:fact ‘Three rounds of soldiers will be surrounding it, you will not be able to (go there and) steal it.’ (2003qachga, 55) (Japhug) The same observation also applies to verbs with AM in complement clauses selected by a verb in the protasis, as in(21): the realization of the verbal action (in addition to that of the motion event) belongs to the condition. (21) nɤʑo ɕɯ-kɤ-mɯrkɯ a-pɯ-tɯ-cʰa nɤ aʑo cʰɯ-sɯ-jɣat-a jɤɣ 2sg go&do-inf-steal irr-ipfv-2-can lnk 1sg ipfv-caus-go.back-1sg be.agreed:fact ‘If you succeed stealing it (after having gone there), I can cause him to go back there.’ (02montagnes-kamnyu, 46) (Japhug) By contrast, in (22), in the case of the ininitival complement kɯ-rɤma kɤ-ɕe ‘go to work’ with a purposive clause (kɯ-rɤma), the main verb mda ‘be time to’ only has scope over the motion event expressed by the verb ɕe ‘go’ – the time that is indicated by the stars refers to the beginning of the journey to work, not the start of the work itself. (22) tɕe kɯɕɯŋgɯ tɕe tɯtsʰot pɯ-me tɕe nɯnɯ cʰɯ-ɬoʁ lnk long.ago lnk clock pst.ipfv-not.exist lnk dem ipfv:downstream-come.out lu-ɕqʰlɤt nɯra ɕ-tu-kɯ-ru tɕe, nɯnɯ kɤ-rɤru ipfv:upstream-disappear dem:pl go&do-ipfv:up-genr:S/P-look lnk dem inf-get.up mda mɤ-mda cʰondɤre kɯ-rɤma kɤ-ɕe mda be.time:fact neg-be.time:fact comit nmlz:S/A-work inf-go be.time:fact mɤ-mda nɯtɕu ɕ-tu-kɯ-ru pɯ-ŋgrɤl. neg-be.time:fact dem:loc go&do-ipfv:up-genr:S/P-look pst.ipfv-be.usually.the.case ‘In former times, there was no clock, and people used to watch when (these stars) came out or disappeared (to know) whether it was time to get up or go to work.’ (29-LAntshAm, 66) (Japhug) 2.4.5 Pluri- vs monoactionality in Situ Minimal pairs similar to those presented above in Japhug have been elicited in Situ. However, it appears that in this language, the use of verbs in perfective form with AM markers does not presuppose that both the verbal action and the motion event have been accomplished. 7 In (23), the action of the verb nə-ɟɐ-ta-natsô-n ‘I came to see you’ with the AM preix ɟɐ- is negated by the following clauses, while the motion event clearly took place; the corresponding Japhug example in § 2.4.1 is considered to be not only agrammatical, but also non-sensical. Example (23) has little discernible semantic diference with the corresponding MVC in (24). rɐ, nəɟə̂ nə-ˈtə-mɐ-n, majnə (23) nə-ɟɐ-ta-natsô-n pfv:downwards-come&do-2→1-lookII -2 conj 2sg sens-2-do.not.existI -2sg conj ma-na-ta-mətɐ̂ -n. neg-pfv-2→1-seeII -2sg ‘I came to see you, but you were not here, so I didn’t see you.’ (Situ) majnə rɐnə, nəɟə̂ nə-ˈtə-mɐ-n, nə-və̄-ŋ (24) nə-kə-natsō poss.2sg-nmlz-see pfv:downwards-comeII -1sg conj 2sg sens-2-do.not.existI -2sg conj ma-na-ta-mətɐ̂ -n. neg-pfv-2→1-seeII -2sg ‘I came to see you, but you were not here, so I didn’t see you.’ (Situ) Similarly, in (25), the verb form nə-ɟɐ-ˈtə-va-n ‘What did you come to do?’ with the AM preix ɟɐdoes not imply that the action of the verb has already taken place, unlike its corresponding Japhug in § 2.4.2. (25) nəɟə̂ tɕɐ̄ thə̂ nə-ɟɐ-ˈtə-va-n? 2sg here what pfv:downwards-come&do-2-doII -2sg ‘What did you come to do?’ (Situ) (26) nəɟə̂ tɕɐ̄ thə̂ kə-viɛ̂ nə-ˈtə-və-n kɐ? 2sg here what inf-do pfv:downwards-2-comeII -2sg part ‘What did you come to do?’ (Situ) This diference between Japhug and Situ, two otherwise relatively closely related languages, suggests that features such as monoactionality of motion event and verbal action may not be diachronically stable. 2.5 Lack of volitionality and control An additional diference between AM and MVC has to do with volitionality and/or controllability. In the case of an MVC, the verb in the purposive clause, whose action follows the motion event, is always necessarily volitional and controllable. By contrast, in the case of AM, it is possible to ind examples where the verbal action expresses a non-controllable event, such as the action of inding a lost object in example (27) with echo phenomenon (cf § 2.3). Note that there are no examples of the non-volitional verb mto ‘see’ with the MVC in the corpus (the volitional rtoʁ ‘see, look’ or ru ‘look’ occur instead). (27) nɯɕɯmɯma ʑo tɯ-ci ɯ-ŋgɯ pjɤ-ɕe qʰe immediately emph indef.poss-water 3sg.poss-inside ifr:down-go lnk iɕqʰa tɤɕime kɯ ɯ-sɤcɯ pɯ-kɤ-nɯ-ɕlɯɣ nɯ the.aforementioned lady erg 3sg.poss-key pfv:down-nmlz:P-auto-drop dem ɕ-pjɤ-mto. go&do-ifr-see ‘He went immediately into the water and saw there the key that the lady had dropped by mistake.’ (140510 fengwang, 118) (Japhug) 8 3 Kiranti The most complex AM systems of the Sino-Tibetan family are found in Kiranti, with systems with up to seven distinct markers (Khaling; Yamphu, Rutgers 1998: 137-194). The main characteristics of AM marking in Kiranti are the following: 1. AM systems of greatly difering complexity across the subgroup 2. Vertical dimension must be described 3. AM markers are labile: they can be either intransitive or transitive (usually matching the main verb, but not always) ; when transitive, the argument of motion is not only the A but also includes the P 4. AM markers often have multiple uses, with the same marker variously also marking aspect, as well as orientation, personal coniguration, and even voice. 5. Verbs with AM marking are pluri-actional (in Khaling; unclear of other languages) In Kiranti, AM markers are ancient auxiliaries (variously called ‘aspectivizers’, ‘bound roots’, ‘motionalizers’, ‘general motion auxiliaries’, ‘V2 s’, ‘vector verbs’, ‘oriented motion verbs’ and ‘modal verbs’ in diferent descriptions of Kiranti languages) which come from motion verbs, though in many cases the source verb does not exist any more. There is evidence that such complex verb forms originate from serial verb constructions, in which a motion verb (as well as other auxiliaries) occupied the second position, but became progressively integrated into a single word (though with diferences across Kiranti; see Bickel et al. 2007, Schiering et al. 2010). For some languages, the second verb has grammaticalized into an aspectual marker (see Bickel 1996), but some of these languages still show, within the examples given in grammars, some situations where the marker clearly indicates AM and not aspect. In other cases, it is not aspect but orientation which is the main function of the markers which can also occasionally mark AM. However, some languages have AM markers which are devoted to motion. These diferent degrees of grammaticalization of auxiliary motion verbs explains one of the interesting features of AM in Kiranti: even with cognate morphemes, the languages range from 0 to 7 AM markers. Example (28) illustrates a typical AM-marked verb in Khaling, with the circumambulative -le‘go around doing’. Note that the 2pl:n.pst suix -ni occurs in reduced form -n- between the stem of the verb -ŋô- and the AM marker -le-, suggesting that at an earlier stage both the main verb and the AM marker had the full array of person indexation and TAM marker. (28) ʔi-ŋô-n-le-ni 2/inv-cry-2pl:n.pst-go.around.doing-2pl:n.pst ‘Youpl went around crying.’ (Khaling) 3.1 Morphosyntactic parameters of AM in Kiranti The Khaling AM system, which ranks among the most complex in the Kiranti branch, can be used to illustrate the general typological characteristics of AM systems in this branch. Table 2 summarizes the data concerning four parameters (temporal relation, deixis, vertical dimension and argument of motion) relating to the AM marker. 9 Source 10 ? |kʰot| ‘go’ ? |ɦo| ‘come’ |kʰoŋ| ‘come up’ |pi| ‘come (same level)’ |tɛn| ‘fall’ V2 Table 2: The Khaling AM system |-le(t)-| ‘go around doing X’ |-kʰo(t)-| ‘do X and go’ |-pɛ(t)-| ‘go and do X’ |-ɦo(t)-| ‘do X and come’ |-kʰoŋ-| ‘do X and come up’ |-pi(t)-| ‘do X and come (same level)’ |-tɛ(nt)-| ‘do X and come/bring down’ Temp. C S P S S S S/P Deixis circum trans trans cis cis cis cis Vertical ∅ ∅ ∅ ∅ up same level down Argument S/A S/A(+P) S/A S/A(+P) S/A S/A(+P) S/A(+P) 3.1.1 Temporal relation As seen in Table 2, AM markers in Kiranti languages can indicate prior, concurrent or subsequent motion (abbreviated as P, C and S respectively) with respect to the main verb. This is exempliied in (29) from Khaling and (30) from Belhare, where the AM markers |-kʰot-| ‘do and go’ and |-thaŋ-| ‘go up and do’ respectively express subsequent and prior motion. (29) tʉ̂ŋ-kʰʌ-tɛ drink-away/do&go-2/3:pst ‘He drank it up’ or ‘He drank it and left.’ (Khaling) (30) cia thuŋ-dhaŋŋ-itt-u! tea set.up.to.cook-go.upwards-accelerative-3U ‘Go up and cook up some tea!’ (Belhare, Bickel 1999: 73) In some cases, pragmatics will trump the usual temporal relation for a given AM marker. This is seen in the set of examples below, where |-tɛ(nt)-| ‘do X and come/bring down’, which generally implies subsequent motion (as in 31), can instead be used to express prior motion if the scenario triggers that interpretation. In (31), there is no obstacle to a default subsequent motion interpretation (buckwheat is cultivated at the same altitude as the villages, and the grain can be beaten without Sirise needing to ascend or descend irst), but in (32), the fact that the rice is cultivated lower in altitude than the household means that only an interpretation whereby Jirise irst descends to the level of the rice ield before beating the rice makes sense to Khaling speakers. (31) siriseʔ-ɛ bʰrɛ̂m rɛp-tɛnd-ʉ p.n.-erg buckwheat beat-come.down-3sg→3.n.pst ‘Sirise beats the buckwheat and comes down.’ (Khaling) (32) dzirise-ʔɛ rɵ̂ː rɛp-tɛnd-ʉ p.n.-erg rice beat-come.down-3sg→3.n.pst ‘Jirise comes down, beats the rice (and returns).’ (Khaling) 3.1.2 Deixis and vertical dimension Most AM systems in Kiranti languages systems have a circumambulative ‘go around doing’ (see (33 from Yakkha), although its form is cognate only in very closely related languages, such as Khaling and Dumi. (33) ŋkha i=ya het-u-ghond-wa-ga? that what=NMLZ.SG cut-3P-V2 roam-npst-2 ‘What are you cutting (at various places)?’ (Yakkha, Schackow 2015: 326) The circumambulative marker is not necessarily always interpreted as AM in all languages; for instance in Belhare, Bickel (1996: 164) shows that the marker which he glosses as ‘spatially distributed temporary’ sdt has both AM and non-AM uses (compare 34 and 35). (34) rot-de i-baŋ-ŋa chap-kon-u road-loc one-hum-obl write-sdt-3u ‘Somebody is walking around taking notes on the road.’ (Belhare) (35) thali khore wat chi-gon plat cup clean clean-sdt ‘He is cleaning plates and cups.’ (Belhare) 11 The traditional deictic categories of cis- and translocative are clearly present in Kiranti languages, and always identiiable through the gloss. With the circumambulative, this results in a three-way deictic system for the languages in the subgroup. Kiranti languages also mark the vertical dimension with cislocative (or more rarely, translocative) deixis (motion down, up, and on the same level), as shown by (36) and (37) from Yamphu. (36) mo.ba ka sem.so semlu-ʔug-iŋ that-ela I sing.too sing-come_down-exps ‘I came down a-singing’ (Yamphu, Rutgers 1998: 143) (37) i.doʔ ik-kætt-u this.like grind-bring_up-3P ‘He ground and brought up [the chutney]’ (Yamphu, Rutgers 1998: 144) Since the AM markers are transparently grammaticalized from these motion verbs (and their applicative/causative counterparts), it is hardly surprising that this contrast is also found in the AM system. Note that a distinction is often made in Kiranti languages between verbs and AM markers for ‘same level’ and ‘unspeciied for level’. 3.1.3 Argument of motion With all AM markers for Kiranti languages, the argument undergoing the motion event always includes the subject (S/A). The AM markers tend, however, to be labile, having both intransitive and transitive forms. When the AM is marked on a transitive verb, in many cases the motion is to be interpreted transitively, resulting in a ‘manipulative’ meaning whereby translocatives are interpreted as ‘take away’ and cislocatives as ‘bring’, the latter exempliied in (38). The result of this lability is that when the AM is transitive, the P is an argument of motion, along with the A . We have indicated this with the formula ‘S/A(+P)’ in tables presenting Kiranti data. (38) eko phuŋ chikt-u-ra=na one lower pluck:pst-3.P-V2 .bring=nmlz.sg ‘She plucked a lower and brought it’ (Yakkha, Schackow 2015: 312) While it is never possible for an intransitive main verb to be accompanied by a transitive version of the AM marker, it must be noted that the AM marker is not always interpreted transitively with a transitive main verb as in (39). (39) ʔīn kʰɵs-tʰer-e uŋʌ ʔʌ-dʌrʌm ɦûŋ-kʰond-u 2sg go-habit-imp:2sg 1sg.erg 1sg.poss-friend wait-do&come.up-1sg.A.n.pst ‘You keep going, I will wait for my friends and come up then.’ [NOT ‘brought them up’] (Khaling) There are a number of examples of an intransitive interpretation for an AM marker on a transitive verb, even though the actual form of the AM marker will in such a case be transitive (in other words, it will be based on a form of the motion verb with an applicative/causative suix). Considering that the interpretation for the transitivity of the AM marker is contextual, this is another instance (as with the temporal relation being afected by context) of default interpretations being trumped by pragmatics. 12 3.1.4 Non-AM meanings As mentioned earlier, what accounts for the diversity of AM systems in Kiranti languages is the fact that diferent languages have undergone diferent degrees of grammaticalization of motion verbs in the second position of what were presumably originally serial verb constructions. In some languages, the motion verbs have all grammaticalized to the point that none can be used to encode AM (this is the case with Bantawa, for example); in other languages, most markers have undergone considerably grammaticalization, but occasionally some uses of the V2 encode AM (an example of this is provided by examples 34 and 35 above from Belhare). In yet other cases, most of the motion verb V2 ’s are used for AM (this is the case of Khaling, where only one AM marker – -kʰot- – is also used to mark completive aspect/telicity, as shown by 29 above, while the others are used exclusively for AM.) The primary non-AM meaning found is aspect. Other non-AM meanings are however found. One of interest is the voice reading found in Hayu, where with transitive verbs, -la(t), which is otherwise an AM marker, is used to express the passive voice as in (40). (40) ma jeŋ la mima jeksa neg voir il-va ainsi nuit ‘Cela ne se verra pas comme ça de nuit.’ (Michailovsky 1988: 153, Hayu) In yet other situations, what are sometimes AM markers can also be used to encode orientation without motion. One of the great diiculties in decoding the examples found in grammars is that AM markers are most often given with verbs of motion or of transfer, where it is not clear whether the marker in question merely encodes orientation or whether it also carries motion. For instance, since the verb |yok| ‘seek’ implies an intrinsic motion, an example such as (41) is not suicient to prove that the V2 -tus- can be analyzed as AM rather than orientation. (41) mo-ba pa:tro yok-tus-iŋ that-ela patro seek-around-exps ‘So then I went around looking for a calendar’ (Rutgers 1998: 153, Yamphu) 3.1.5 Mono- vs pluriactionality As discussed for Gyalrongic languages, the mono- vs pluri-actionality of the verb+AM unit is interesting to consider in Kiranti. Again, one of the diiculties in broadening this discussion to Kiranti as a whole results from the fact that it is rarely possible to determine the extent of pluri-actionality from the examples found in many grammars. Nonetheless the irst-hand data we have collected on Khaling suggests that AM markers form pluri-actional verb events: when examples are negated, the scope of the negation is limited to the main event and does not cover the AM. This can be seen in example (42) (compare with 39 above). (42) ʔuŋʌ ʔʌ̄m mʌ-ɦû-n-kʰōː-nɛ-ʔɛ ʔu-nûː ŋes-tɛ 1sg.erg 3sg neg-wait-inf-do&come.up-inf-erg 3sg.poss-mind hurt-2/3:pst ‘I did not wait for him before going up and he was sad. (I was already gone up by the time he arrived at the waiting place).’ (Khaling) The pluri-actionality of the verb+AM in Khaling has only been tested with negation, but presumably the tests described for Gyalrongic languages would work equally well with Kiranti languages. 13 3.2 Survey of AM systems in Kiranti A large number of grammars of Kiranti languages have been consulted for the present survey, but not all languages have been included, as shorter grammars (Ebert 1997b, Ebert 1997a, Opgenort 2005, Tolsma 2006) do not always include enough data on V2 . Not all Kiranti languages have AM. Although Bantawa (Doornenbal 2009) is among the better described languages, no example of V2 clearly analyzable as AM could be found in the available publications. The simplest AM system is that of Thulung (Lahaussois 2002 and additional ieldwork) with only one AM marker, the circumambulative -bal (table 3). Dumi, the closest relative of Khaling, has at least the four markers indicated in Table 4, possibly more (van Driem 1993: 199-214). Note that all the V2 in this Table have Khaling cognates, with similar functions, and that the AM system in these two languages may go back to their common ancestor. Wambule (Opgenort 2004) has ive AM markers, all of which can have a manipulative ‘bring/ take’ meaning (Table 5). The verb form (43) with -lwa- ‘go/take and do’ provides a clear example of the AM value of this V2 . (43) sei-lwa-s-ta kill.self-go/take-detr-imp:sAS ‘go and kill yourself’ (Opgenort 2004: 439) (Wambule) Yamphu has a particularly rich AM system, with at least seven markers (Rutgers 1998: 137-194). Among the V2 listed as AM markers in table 6, note that -phæt(t)- essentially has non-AM functions, but does occur with the meaning ‘do and leave’ as in example (44), hence its analysis as a subsequent translocative AM marker. (44) uŋ-bhæ:tt-o uŋ-bhæ:tt-o sapphi drink-away-arq drink-away-arq abundantly ‘Drink before you leave, drink before you leave, [drink] as much as you can.’ (Rutgers 1998: 150). (Yamphu) The Yamphu V2 -las- ‘go, do and come back’ implies two motion events, and may be similar to the ‘interrupted motion’ category proposed by Rose (2015: 123) to describe ‘situation where the realization of the lexical event occurs between two stretches of motion and where the motion is encoded by a marker distinct from that of ‘concurrent motion’. 14 Source |bal| ‘wind around’ V2 Table 3: The Thulung AM system -bal- ‘go around doing’ 15 -pad- ‘to go of to do something’ -li(lɨt)- ‘to be up and about while doing’ -hu:d- ‘do and bring back here’ -pid- ‘do and bring it over here’ |pi| ‘come (horizontal)’ |phit| ‘bring (horizontal)’ |ga| ‘come up’ |khat| ‘bring up’ |ywa| ‘come down’ |hywat| ‘bring down’ |di| ‘go and come back’ |lwa| ‘go’ |lyat| ‘take away’ C Deix. Vert. circum ∅ Arg. S/A+P Table 4: The Dumi AM system (van Driem 1993: 199-214) V2 Source Temp. V2 Gloss Allative Frolicsome fetch bring Temp. P C S S Deix. trans circum cis cis Table 5: The Wambule AM system (Opgenort 2004) ∅ ∅ ∅ ∅ Arg. S/A S/A S/A+P S/A+P come/bring:hrz Temp. P Deix. Vert. same level S/A+P -kha- ‘come/bring up’ ? P cisl up S/A+P -ywa,hywa- ‘come/bring down’ come/bring:down P/I cisl down S/A+P -di,du- ‘go/take (and come back)’ -lwa- ‘go/take and do’ go/take/come go/take P P transl cisl ∅ same level S/A+P S/A+P -phi- ‘come/bring and do’ Gloss Vert. cisl Arg. Source V2 |ap| ‘come (horizontal)’ |uks| ‘come down’ |kæt| ‘come up’ |las| ‘go and come back’ Source Table 6: The Yamphu AM system (Rutgers 1998: 137-194). -ca/cæt- ‘come and do’ -ap(t)- ‘come and do’ -uk(t)- ‘come down and do’ -kæt(t)- ‘do and come/bring up’ -phæt(t)- ‘do and leave’ -las- ‘go, do and come back’ -tus/tit- ‘go around doing’ Temp. come/bring come/bring come/bring down come/bring up away go come around Deix. P P P/C S S S+P C cisl cisl cisl cisl transloc cisl circum Table 7: The Yakkha AM system (Schackow 2015: 283-328). V2 16 -kheʔ/t- ‘do and go/carry’ -uks/t- ‘do and come/bring down’ -ap(t)- ‘do and come/bring’ -ghond- ‘go around doing’ Source Gloss V2 Gloss V2 .go/carry.off V2 .come.down/bring.down V2 .come/bring V2 .roam Temp. Deix. S S S C trans cis cis circum Table 8: The Belhare AM system (Bickel 1996, 1997, 2017). -itt- ‘go up and do’ -ap(t)- ‘do and come/bring’ -kon- ‘go around doing’ Gloss go.upwards bring.across spatially distributed temporary Temp. S P C Deix. cis trans circum Table 9: The Hayu AM system (Michailovsky 1988: 151 Source |lat| ‘go’ V2 -la(t)- ‘go and do’ Temp. P Deix. trans Vert. ∅ Arg. S/A Vert. Arg. ∅ same level same level same level ∅ ∅ ∅ Vert. Arg. ∅ down same level ∅ Vert. down up ∅ S/A+P S/A+P S/A+P S/A+P S/A S/A+P S/A S/A, A+P S/A+P S/A+P S/A Arg. S/A+P S/A+P S/A 4 Sinitic AM has been rarely addressed in linguistic studies of the Sinitic languages. It has not been described until in Lamarre (Forthcoming) and Lamarre et al. (Forthcoming) where the sequences [VP+去/来 qu/lai] ‘go/come to do’ in Northern Sinitic languages are analyzed as am encoding. In (45) and (46), the andative enclitic 去 qu and the ventive enclitic 来 lai respectively encode a translocative and a cislocative motion in addition to the action denoted by the VP. (45) Standard Mandarin (Lamarre et al. Forthcoming) 点儿 水=去 喝 Hē diǎnr shuǐ =qu drink some water=go& ‘Go and drink some water!’. (46) Standard Mandarin (Lamarre Forthcoming) 了 你 干 嘛=来 le Nǐ gàn má=lai 2sg do what=come& part ‘What are you coming for?’ Note that unlike Gyalrongic and Kiranti, the host of the AM marker is not the verb itself, but the verb phrase, and occurs after the object as in (45) and (46). Its is thus a borderline case of AM; note that Guillaume (2016), which is restricted to aixal AM markers, does not include markers that are clitics. However, it does conform to Guillaume’s (2016: 13) deinition of AM as ‘a grammatical morpheme that is associated with the verb and that has among its possible functions the coding of translational motion’, and given the problematic and controversial nature of word boundaries in Sinitic, it is unsurprising that these languages lack clearly aixal AM markers. The [VP+去/来 qu/lai] construction was previously described as a type of ‘purpose construction’ (Lu 1985; Yang 2012), the inal 去/来 qu/lai analyzed as ‘particles of purpose’ (Chao 1968: 479). The construction is considered to be interchangeable to some degree with the serial verb construction (henceforth SVC) [去/来 qu/lai+VP]. We can compare (45) and (46) with (47) and (48) below. Noted that as am markers (in 45 and 46), the post-verbal 去/来 qu/lai lose their verbality to some degree and show tendency to become bound morphemes, with phonetic weakening compared to their source verbs (in 47 and 48). Chao 1968: 479; Lu 1985; Lamarre (Forthcoming) ; Lamarre et al. (Forthcoming)). (47) Standard Mandarin (Lamarre et al. Forthcoming) 去 喝 点儿 水 Qù hē dianr shuǐ go drink some water ‘Go and drink some water!’ As shown by (46) and (48), AM markers in Mandarin allow pluri-actional reading, unlike those found in Japhug (see § 2.4.2). (48) Standard Mandarin 17 你 来 干 什么 Nǐ lái gàn shénme 2sg come do what ‘What are you coming for?’ Source 去 qù 来 lái Table 10: am markers in Mandarin Enclitic Temp. Deix. Vert. =qu ‘go and X’ =lai ‘come and X’ P P ∅ ∅ ∅ ∅ Arg. S/A S/A No Sinitic language has dedicated AM markers. Depending on the variety, AM enclitics can also mark orientation (in particular when used with motion verbs), aspect and imperative. 5 Other languages The term ‘AM’ has rarely been used to describe the grammar of ST languages other than Gyalrongic, Kiranti and Sinitic – to our knowledge, only the recent grammars of Karbi (Konnerth (2014)) and Hakhun Tangsa (Boro (2017)) make use of this term. 5.1 Karbi Konnerth (2014, 2015) reports the existence of a cislocative AM proclitic nang= in Karbi, whose use can be exempliied by (49). nang=vùr-si nang=jùn-lò (49) alàng-lì=ke là-tūm a-hēm=si sá aját 3-hon=top this-pl poss-house=foc cis=drop.in-nf:rl tea genex cis=drink-rl ‘...it was him, at their house we stopped by and had tea [come and drink tea] and everything.’ In addition, this proclitic occurs as a person indexation marker, used for irst or second person nonsubject (including local scenarios 1→2 and 2→1)as in (50) and as a marker of cislocative orientation with (51) or without motion (52). (50) nang-phān=ke nang=kV-pòn-pò you-nsubj=top 1/2:nsubj=ipfv-take.away-irr1 ‘[…] (I) will carry you away’ (51) lasō a-hūt amāt [e-nūt a-kV-prék a-monít abàng=ke] this poss-during and.then one-clf:hum.sg poss-nmlz-be.diferent poss-man npdl=top saikél nang=ardòn-si vàng-lò… bicycle(<Eng) cis=ride-nf:rl come-rl ‘in this moment, another person riding on a bicycle came, (52) angsóng=pen=si phén nang=jāng-lìng up.high=from=foc fan(<Eng) cis=hang-small:s ‘the fan is hanging down from up high (from the ceiling)’ (Karbi) 18 5.2 Hakhun Tangsa Boro (2017: 311-312) analyses the construction in (53) as AM. It remains however unclear from the description whether an alternative analysis as MVC is not possible. (53) inɤ́ ʒuk kà l-oʔ there drink go imp-2sg ‘Go and drink (tea) there.’ (Hakhun Tangsa) 5.3 Burmese Burmese has four preverbal (la ‘come’, θwà ‘go’, laiʔ ‘follow, go around and’, cauʔ ‘walk, go around and’) and two postverbal auxiliaries (-la ‘come’ and -θwà/ðwà ‘go’) that can be used to express a motion event (Jenny & Hnin Tun 2016: 185-190;203-5) and are candidates to be analyzed as AM. Preverbal auxiliaries are used for prior and concurrent motion events (54) and postverbal ones to express subsequent motion (55). (54) θu ywa-hma paiʔsʰan laiʔ-tàun-dɛ 3 village-at money follow-ask.for-nfut ‘He went around the village begging for money’ (Jenny & Hnin Tun 2016: 189, Burmese). (55) θu-dó tʰəmìn sà-la-dɛ 3-pl cooked.rice eat-come-nfut ‘They have eaten (before coming here).’ (Jenny & Hnin Tun (2016: 388), Burmese) 6 Summary Subgroup Language Number Temp. relation Deixis Gyalrong Japhug Situ 2 2 P P Kiranti Khaling 7 Dumi Sinitic Karbi Sal LB Vertical dimension Other functions Monoactional Arg. cis/trans cis/trans ∅ ∅ ∅ Aspect ! % S/A S/A P, C, S cis/trans/circum ! % S/A, S/A+P 4? P, C, S cis/trans/circum ∅ ? S/A, S/A+P Thulung Wambule 1 5 C P circum cis/trans/circum ∅ ! ? ? S/A S/A, S/A+P Yamphu 7 P, C, S cis/trans/circum ! ? S/A, S/A+P Yakkha 7? P, C, S cis/trans/circum ! ? S/A, S/A+P Belhare 3? P, C, S cis/trans/circum ! ? S/A, S/A+P Hayu 1 P trans ∅ Aspect, Orientation Aspect, Orientation ∅ Aspect, Orientation Aspect, Orientation Aspect, Orientation Aspect, Orientation Aspect, Voice ? S/A Mandarin 2 P cis/trans ∅ % S/A 1 S cis ∅ Tangsa Burmese 1 6? P P, C, S trans ∅ cis/trans/circum 19 ∅ Aspect, Modality Orientation Person, Orientation ? Aspect, Orientation ? ? ? S/A +P S/A S/A, S/A+P 7 Conclusion AM systems are by no means rare in the ST family; in particular, the presence of AM in Mandarin and Burmese, the two languages with the largest number of speakers, suggest that it is possibly a very widespread if under-reported phenomenon. AM systems have been grammaticalized from motion verbs independently several times in the ST family, and present considerable typological diversity. Each of the three subgroups discussed in detail in this work contributes to the typology of AM in a diferent way. Gyalrong languages have simple AM systems with only two members, but the high frequency of the markers in texts makes it possible to describe in greater detail the semantic properties of AM in these languages. A corpus study on Japhug shows in particular that AM markers difer in many ways from the corresponding purposive construction with motion verbs (MVC): several syntactic tests show that the motion event cannot be separated from the action of the main verbs in the case of AM. Moreover, AM can be used with non-volitional verbs, unlike MVC. Kiranti languages have richer AM systems, but their AM markers are less common in texts, and lack echo phenomena (unlike Gyalrong, 2.3) and monoactionality. They do however encode more parameters, in particular the temporal relation (AM markers for prior, concurrent and subsequent motion event are found), the vertical dimension and also include manipulative AM markers (‘bring/ take and X’). The languages with the highest number of AM markers in the ST family (Khaling and Yamphu) are found in this branch. Sinitic languages have borderline cases of AM systems, as AM markers in these languages are clitics attached on the verb phrase rather than verbal aixes. The diversity of this branch and the existence of continuous written attestations potentially allows to follow in greater detail the grammaticalization process of AM. The study of AM in this branch however ofers particular challenged, in particular distinguishing between AM and directional complements. The present work contributes to the typology of AM by presenting in more detail a range of tests and criteria that can be used to reine the study of the semantic contrast between AM and other constructions. Previous descriptive work on AM in ST, even in the most detailed grammars, rarely provides explicit examples making it possible to determine whether pluriactional interpretation of the motion event + action is possible. The sharp contrast between Japhug on the one hand, and Khaling with regards to this parameter shows that ieldworkers should devote speciic attention to this providing examples which make clear the presence or absence of this phenomenon. References Alonso de la Fuente, José Andrés & Guillaume Jacques. to appear. Associated motion in Manchu in typological perspective. Language and Linguistics . Bickel, Balthasar. 1996. Aspect, mood, and time in Belhare: studies in the semantics-pragmatics interface of a Himalayan language: Universität Zürich dissertation. Bickel, Balthasar. 1997. Spatial operations in deixis, cognition, and culture: where to orient oneself in Belhare. In Jan Nuyts & Eric Pederson (eds.), Language and conceptualization, 46–83. Cambridge: Cambridge University Press. Bickel, Balthasar. 1999. Cultural formalism and spatial language in Belhara. In Balthasar Bickel & Martin Gaenszle (eds.), Himalayan space: cultural horizons and practices, 73–101. Zürich: Museum of Ethnography. Bickel, Balthasar. 2017. Belhare. In Graham Thurgood & Randy LaPolla (eds.), The Sino-Tibetan Languages, Second edition, 696–721. London: Routledge. 20 Bickel, Balthasar, Goma Banjade, Martin Gaenszle, Elena Lieven, Netra Paudyal, Ichchha Purna Rai, Manoj Rai, Novel Kishore Rai & Sabine Stoll. 2007. Free preix ordering in Chintang. Language 83(1). 1–31. Boro, Krishna. 2017. A Grammar of Hakhun Tangsa: University of Oregon dissertation. Chao, Yuanren. 1968. A Grammar of Spoken Chinese. Berkeley: University of California Press. Doornenbal, Marius. 2009. A Grammar of Bantawa: Grammar, paradigm tables, glossary and texts of a Rai language of Eastern Nepal. Leiden: Leiden University dissertation. van Driem, George. 1993. A Grammar of Dumi. Berlin and New York: Mouton De Gruyter. Ebert, Karen H. 1997a. Camling. München: Lincom Europa. Ebert, Karen H. 1997b. Grammar of Athpare. München: Lincom Europa. Gong, Xun. 2018. Le rgyalrong zbu, une langue tibéto-birmane de Chine du Sud-ouest : une étude descriptive, typologique et comparative: Institut national des langues et civilisations orientales dissertation. Guillaume, Antoine. 2016. Associated motion in South America: Typological and areal perspectives. Linguistic Typology 20(1). Jacques, Guillaume. 2013. Harmonization and disharmonization of aix ordering and basic word order. Linguistic Typology 17(2). 187–217. Jacques, Guillaume. 2016. Complementation in Japhug. Linguistics of the Tibeto Burman Area 39(2). 222–281. Jenny, Mathias & San San Hnin Tun. 2016. Burmese, a comprehensive grammar. London: Routledge. Koch, Harold. 1984. The category of ‘Associated Motion’ in Kaytej. Language in Central Australia 1. 23–34. Konnerth, Linda. 2014. A grammar of Karbi: University of Oregon dissertation. Konnerth, Linda. 2015. A new type of convergence at the deictic center: Second person and cislocative in Karbi (Tibeto-Burman). Studies in Language 39(1). 24–45. Lahaussois, Aimée. 2002. Aspects of the grammar of Thulung Rai an endangered Himalayan language: University of California, Berkeley dissertation. Lai, Yunfan. 2017. Grammaire du khroskyabs de Wobzi: Université Paris III dissertation. Lamarre, Christine. Forthcoming. An Associated Motion approach to northern Mandarin MOTIONCUM-PURPOSE patterns, . Lamarre, Christine, Voisin Kopecka & al. Forthcoming. Deictic directionals revisited in the light of advances in typology . Lin, Youjing. 2003. Tense and aspect morphology in the Zhuokeji rGyalrong verb. Cahiers de linguistique - Asie orientale 32(2). 245–286. Lu, Jianming 陆俭明. 1985. Guanyu qu VP he VP qu jushi [On the patterns ‘qu VP’ and ‘VP qu’]. 语 言教学和研究 4. 18–33. Michailovsky, Boyd. 1988. La langue hayu. Paris: Editions du CNRS. 21 Michailovsky, Boyd, Martine Mazaudon, Alexis Michaud, Séverine Guillaume, Alexandre François & Evangelia Adamou. 2014. Documenting and researching endangered languages: the Pangloss Collection. Language Documentation and Conservation 8. 119–135. Opgenort, Jean-Robert. 2004. A Grammar of Wambule. Leiden: Brill. Opgenort, Jean-Robert. 2005. A grammar of Jero, with a historical comparative study of the Kiranti languages. Leiden: Brill. Prins, Marielle. 2016. A grammar of rGyalrong Jiaomuzu (Kyom-kyo) dialects, A web of relations. Leiden: Brill. Rose, Françoise. 2015. Associated motion in Mojeño Trinitario: Some typological considerations. Folia Linguistica 49(1). 117–158. Rutgers, Roland. 1998. Yamphu, Grammar, Texts and Lexicon. Leiden: Research School CNWS. Schackow, Diana. 2015. A grammar of Yakkha Studies in Diversity Linguistics 7. Berlin: Language Science Press. http://langsci-press.org/catalog/book/66. Schiering, René, Balthasar Bickel & Kristine A. Hildebrandt. 2010. The prosodic word is not universal (but emergent). Journal of Linguistics 46. 657–709. Stoynova, Natalia M. 2016. Pokazateli “dvizhenija s cel’ju” i sobytijnaja struktura: suiks -nda v nanajskom jazyke. Voprosy Jazykoznanija (4). 86–111. Sun, Jackson T.-S. 2012. Complementation in Caodeng rGyalrong. Language and Linguistics 13(3). 471–498. Sun, Jackson T.-S. 2014. Sino-Tibetan: Rgyalrong. In Rochelle Lieber & Pavol Štekauer (eds.), The Oxford Handbook of Derivational Morphology, 630–650. Oxford: Oxford University Press. Tolsma, Gerard Jacobus. 2006. A Grammar of Kulung. Leiden: Brill. Vuillermet, Marine. 2012. A grammar of Ese Ejja, a Takanan language of the Bolivian Amazon: Université Lyon II dissertation. Wilkins, David P. 1991. The Semantics, Pragmatics and Diachronic Development of ‘Associated Motion’ in Mparntwe Arrernte. Bufalo Papers in Linguistics 1. 207–57. Yang, Yonglong 杨永龙. 2012. The purpose construction ‘VP qu’ in Chinese and SOV order 目的 式‘VP去’与SOV的语序关连. Zhongguo Yuwen 中国语文 (2). 525–576. Zhang, Shuya. 2016. La phonologie et la morphologie du dialecte de Brag-dbar du rgyalrong situ. Université Paris III MA thesis. 22