# Verb-initial grammars: A multilingual/parallel perspective # ESRC Project RES-000-23-0505 # Oxford University # Charles Randriamasimanana # Malagasy phonetics/phonology & morphology: # MORPHOLOGICAL SEGMENTATION: ESRC-OX-05-CR102. # Malagasy is rich with words, particles or morphemes, which at first sight look very similar, # but which upon analysis behave rather differently. One typical example involves 'ho', which may # on the one hand be a complementizer as in the first instance of 'ho' in the following # illustrative example adapted from a Malagasy online newspaper; or which on the other hand, # may also be a passive voice future tense-marker, as in its second occurrence below: Nambarany fa tsy maintsy hiteraka korontana koa izany aorian’ny fifidianana, # N-ambara-ny fa tsy maintsy h-i-teraka korontana koa izany aorian(a)’ny fifidianana, # past-root.ambara-clitic.him/her comp must fut-pref.i-root.teraka turmoil also that after the election # "Was declared by him that will necessarily cause turmoil that after the election" # "He declared that this will necessarily cause turmoil after the eclection" satria tsy mitovy zo ireo ho kandidà, indrindra eo amin’ny hoenti-manana. # satria tsy m-itovy zo ireo ho kandidà, indrindra eo amin’ ny ho-enti(na)-manana. # since not be.same right those comp candidates, especially there with the pass.fut-stem.enti-to.have # "since not of same right candidates, especially as far as how to get (this) done" # "as those who will be candidates do not have the same rights, especially as far as the resources are # concerned." # Adapted from Malagasy language online newspaper http://taratramada.com/ dated 14/1/2006. # Note that as a complementizer 'ho' tends to remain separate from its non-verbal predicate (either an adjective # or a noun), whereas as a passive voice future tense-marker it is attached to the stem of the verb as it represents # a circumfix made up of a prefix 'ho' and a suffix '(i)na'. # In addition, whether 'ho' is a complementizer or a tense-marker, it is rather restricted as to what other elements # it can co-occur with; in other words, its function depends on the other elements with which it co-occurs. # As a first approximation, the following examples from Paul et al. (1998:55) do not observe the relevant restrictions, # hence the ungrammaticality of many of the sequences immediately below: # Ungrammatical: Mihevitra zanaka ho hendry aho. (0! 0 0 0) # mihevitra zanaka ho hendry aho. (0! 0 0 0) # m-i-hevitra zanaka ho hendry aho # pres-prf-root-think child(ren) comp wise 1 sg(nom) # 'I think some child is well behaved'; # the phrase 'zanaka' must be definite; Mihevitra ny zanako ho hendry aho. # mihevitra ny zanako ho hendry aho. # m-i-hevitra Det-unspec zana-ko ho hendry aho. # pres-prf-root-think the child-of-mine comp wise I # 'I believe my child to be wise' # the phrase 'ny zanako' is definite and the sequence is grammatical; # Ungrammatical: Mihevitra tena ho mahay i Paoly. (0! 0 0 0) # mihevitra tena ho mahay i Paoly. (0! 0 0 0) # m-i-hevitra tena ho ma-hay i Paoly. # pres-prf-root-think self comp-ho prf-intelligent D-sg Paul # comp-ho STATE # 'Paul considers her/himself intelligent' # the reflexive 'tena' must be incorporated into the verb ending in 'tra'; Mihevi-tena ho mahay i Paoly. # mihevi-tena ho mahay i Paoly. # m-i-hevi-tena ho ma-hay i Paoly. # pres-prf-root-think-self comp prf-intelligent D-sg Paul # comp-ho STATE # 'Paul considers her/himself intelligent' # once the relevant adjustment has been made, the sequence is grammatical. # 1 - A common complication which arises in this language is that one segment, i.e. 'na' can represent # totally different morphemes related to suffix 'ina' of passive 1 --promotion of a Direct Object to # Subject-- for instance, or to an unrelated epenthetic 'na'. In order to determine the function of # segment 'na' therefore, one needs to look at the exact nature of the environment in which it occurs: # when 'na' occurs with a stem, then it is more than likely to be part of one of three types of # construction, which entails a DELIBERATE kind of ACTIVITY: # Construction I: circumfix no/ho... ina 'passive.voice.past/future + stem + suffix.ina' # with its variant ø... ina 'passive.voice.present.tense + stem + suffix.ina'; # see the case of 'karamaina' 'is being employed' or 'is usually employed' immediately below; or # Construction II: circumfix an...ana 'passive.2 + stem + suffix ana --promotion of an Instrument-Oblique # to Subject; see Case A & Case C described under 'Malagasy syntax/semantics: Malagasy Prepositions - # ESRC-OX-05-CR207; Malagasy Prepositions.testfile'. # or even Construction III: circumfix aha...ana 'passive.2 + stem + suffix ana --promotion of a # Circumstantial-Oblique to Subject; see illustrative examples with verbs like 'tonga' V '(have)-arrived' # under 'Malagasy lexicon: Malagasy Roots with a passive meaning: ESRC-OX-05-CR301.'; # when by contrast 'na' occurs with a root, then it is more than likely to correlate with a STATE or a SPONTANEOUS/ # ACCIDENTAL kind of EVENT; see for example the case of 'lena' root Adj '(be)soaked' imediately below; # passive.1 suffix ina (or sometimes ana) contrasted with epenthetic or clitic na: # even though suffix 'ina' may take a shape orthographically similar to epenthetic 'na' # when the initial vowel drops out in certain environments, the two have rather different # distributions since suffix 'ina' can only combine with a stem, whereas epenthetic 'na' # can only (usually) accompany a root; furthermore, on a stem stress is almost always one # syllable to the right compared to its location on a root (where non-monosyllabic stems # are concerned). # passive.1 suffix ina combines with a stem, NOT a root & refers to DELIBERATE kind of ACTIVITY: karamaina # ø- karama- ina # passive.present.tense-stem.karama-suffix.ina # Either 'is being employed' or 'is usually employed'; # note the emergence of the zero tense-morpheme for passive present; # this ø zero morpheme is part of the circumfix ø...ina; # the relevant base of the word is the stem karama, with stress on 'ma' # if we were dealing with the root, the latter would have its stress on 'ra'; # use of the stem instead of the root points to a DELIBERATE kind of ACTIVITY; # 'ø' morpheme is translated either as "is usually paid a salary" or "is being # paid a salary"; nokaramaina # no- karama- ina # passive.past.tense-stem.karama-suffix.ina #'Was employed' # prefix no is part of the circumfix no...ina; # the relevant base of the word is the stem karama, with stress on 'ma' # if we were dealing with the root, the latter would have its stress on 'ra'; # use of the stem instead of the root points to a DELIBERATE kind of ACTIVITY; # 'no' morpheme is translated as "was paid a salary", with PUNCTUAL aspect; hokaramaina # ho-karama-ina #'Will be employed' # prefix ho is part of the circumfix ho...ina; # the relevant base of the word is the stem karama, with stress on 'ma' # if we were dealing with the root, the latter would have its stress on 'ra'; # use of the stem instead of the root points to a DELIBERATE kind of ACTIVITY; # 'ho' morpheme is translated as "will be paid a salary", with PUNCTUAL aspect; # epenthetic 'na' combines with a root, NOT a stem & refers to a STATE, and does not # describe a DELIBERATE kind of ACTIVITY : lena # root Adj '(be)soaked'; lenan'i Paoly ny fiara # lena-n' i Paoly ny fiara # root.lena-ep.na D.sg Paul Det car # literally:"The car was accidentally soaked by Paul." # note the use of the root Adj lena '(be)soaked', with final dental 'n'; lemo # '(let-sth)-be-soaked-by.you'; # passive 1 imperative lemo lem+u 'stem.lem-suffix.u' > lem+u # '(let-sth)-be-soaked-by.you (deliberately/intentionally)'; # note the monosyllabic stem 'lem', with final 'm' instead final 'n'; leman'i Paoly ny fiara # ø-lem-an(a) i Paoly ny fiara # passive.present.tense-stem.lem-suff.(a)na D.sg Paul Det car # literally:"The car was deliberately soaked by Paul" # note that in this case, we have circumfix 'ø...ana', i.e. stem 'lem' combining # with passive voice affixes 'ø ... ana'; # 2- incorporation of possessor-genitives into complex DetPs of words # ending in {ka, tra, na} ny alikan'i Jeanne # ny alika-n' i Jeanne # Det-unspec alika-of D-sg Jeanne # the dog(s)-of Jeanne # "Jeanne's dog(s)"; # note the addition of the epenthetic n' on the headnoun indicating that # 'i Jeanne' is in the genitive case; ny menak'omby # ny menak' omby # Det-unspec grease-of beef # the grease of beef # 'the grease of beef'; # note the elision of the final vowel 'a' of the word 'menaka''oil' # indicating that 'omby' is in the genitive case; ny hevitr'olona # ny hevitr' olona # Det-unspec idea(s)-of a/person(s) # 'the ideas of people'; # aagain note the elision of the final vowel 'a' of the word 'hevitra''idea(s)' # indicating that 'olona' is in the genitive case; ny hevitry ny olona # ny hevitry ny olona # Det-unspec idea(s)-of Det-unspec person(s) # 'the ideas of the people'; # note that unlike the earlier cases where the genitive was indefinite, # here the genitive is definite as indicated by the presence of the definite # article 'ny'; some kind of agreement obtains between the final vowel of 'tra' # of 'hevitra' and the article 'ny'; ny fifaneken-dry Paoly # ny fifanekena - n - ry Paoly # Det-unspec contract -of - D-pl Paul # the contract of Paul & associates; # where fifanekena f+if+an+ekena 'nominaliser.f-reciprocal.if-prefix.an-root.ekena', # ekena is a root verb meaning 'be-accepted'; note that 'dry Paoly' is in the genitive # case as indicated by the loss of the final vowel 'a' of the deverbal noun fifanekena; ny fifanekena # ny f -if -an -ekena # Det-unspec nprf-rcp-caus-root-be-accepted # 'the contract between X and Y'; ny fifanarahan'olon-tsotra # ny fifanarahana + n + olona tsotra # Det-unspec agreement of person(s) ordinary # ny f -if -an -arah-an(a) # root araka # Det-unspec nprf-rcp-caus-root-be-followed-passive # 'the agreement between ordinary people'; # where fifanarahana f+if+an+arah+ana 'nominaliser.f-reciprocal.if-prefix.an-stem.arah # -suffix.ana'and where stem arah derives from the root araka 'be-able-to-be-followed'; # note the apostrophe on the final syllable of fifanarahan' indicating a genitive case # on 'olon-tsotra'; olon-tsotra is a combination of two words, i.e. olona 'people' and # tsotra 'ordinary'; # BUT ny famonoana olona # ny famonoana olona # Det-unspec killing people # the killing-of person(s) # ny f+a(n>m)+(v)ono+an(a) 'Det-unspec nominaliser.f-caus.prf.an-root.vono-passive. # suffix.ana' > ny f+am+ono+ana 'the killing of person(s)'; # note the presence of the final vowel a of 'famonoana' despite the subsequent initial # vowel of the word 'olona'; the word 'olona' is a direct object of the deverbal noun; # 3- Nominalization with HABITUAL prefix f on the root form of the verb (NOT the passive 2 # /circumstantial/relative form of the verb): fiteny # 'something which can usually/habitually be said'; # f - i -teny # nprefix f - prf i -root teny # note that in this case, we have f+i+teny 'nominaliser.f-prefix.i-root,teny' as opposed # to the otherpossibility, fitenenana f+i+tenen+ana 'nominaliser.f-prefix.i-stem.tenen- # suffix.ana''the act of saying something'; note that itenenana is the passive 2 form of # the verb m+i+teny 'pres-prefix.i-root.teny''to speak'; # note the presence within the word 'fiteny' of the root teny, whereas in 'fitenenana' # we find the passive 2 stem itenenana; in the on-line document entitled 'Nominalization # in Malagasy', we will distinguish between Stative Nominalization correlating with root # as contrasted with Action Nominalization involving a stem. fifanekena # 'mutual agreement'; # f + if + an + ekana = nominalizer prefix f + reciprocal if + lexical # causative an + root passive ekena 'be accepted' (see above); note the presence of the stem # form ekena; # 4- Nominalization with AGENT nominal prefix mp(a) on the stem of a control verb: mpandeha # 'goer' # mp-andeha (but see below) # nprf-go # m-andeha 'pres-go'; # note that the nominal AGENT prefix mp can be affixed to a verbal predicate like mandeha # where the EXPERIENCER retains control over whether s/he will go or not; mpitaiza # 'minder' # mp -i -taiza # nprf - prf i -root taiza 'offspring' # m-itaiza 'pres-mind'; # note that the nominal AGENT prefix mp can be affixed to a verbal predicate # like mitaiza where the AGENT retains full control over whether s/he does # the minding or not; note the root form taiza 'offspring'. # Ungrammatical: mpanantena (0! 0 0 0) # 'the-one(s)-who-hope'; # mpanantena mp+an+antena 'prefix.mp-prf.an-root.antena'; where root antena = # adverb 'for oneself' and where the notion of control is not necessarily relevant # at least without a proper, somewhat elaborate and explicit discourse context! # There exist in Malagasy at least two tests to show that a given predicate involves # control or not; the first has to do with the possibility of using a control verb # within the framework eo amp + passive 2 verb 'there.visible+prefix.amp+passive 2' # --see Randriamasimanana (1999:518); this test yields eo ampanantenana (0! 0 0 0) # 'being in the middle of expecting'; # eo amp-an-anten-ana 'there.visible+prefix.amp+prefix.an+stem.anten+suffix.ana', # where the combination an-anten-ana represents the passive 2 form corresponding to # the verb manantena 'to expect'; # the second test allows a control predicate to be used within this other framework # -see Randriamasimanana (1986:69-70) # fanahy inian' DP (definite phrase) ny S 'sould willed-by DP comp S fanahy inian'i Paoly ny manantena (0! 0 0 0) # 'Paul is deliberately expecting'; # On the other hand, the same tests on mandeha 'pres-go' yield: eo ampandehanana # 'in the middle of going'; # eo amp+an+(l>d)eha+n+ana 'there.visible-prefix.amp-prefix.an-root.leha-suffix.ana'; # alternatively eo amp+andehanana 'there.visbile-prefix.amp-passive2.andehanana'; fanahy inian'i Paoly ny mandeha # 'Paul deliberately goes'; # fanahy inian' i Paoly ny m-andeha # soul willed-by D.sg Paul comp pres-go # 'Paul deliberately usually goes'; # Likewise, the same tests on mitaiza 'pres-mind'; eo ampitaizana # 'in the middle of minding'; # eo amp+i+taiza+ana 'there.visible-prefix.amp-prefix.i-root.taiza-suffix.ana' or # alternatively eo amp+itaizana 'there.visible-prefix.amp-passive2.itaizana'; fanahy inian'i Paoly ny mitaiza # 'Paul deliberately usually minds'; # fanahy inian'i Paoly ny m-i-taiza # soul willed-by D.sg Paul comp pres-prf.i-mind # 5- Reciprocal infix 'if' restricted to combining with lower CONTROL verbs, i.e. # productive causative prefix an, which can accommodate a whole range of verbal # constructions fifanekena # Ungrammatical: fifekena (0! 0 0 0) # 'mutual agreement'; # f + if + an + ekana 'nominaliser.f-reciprocal.if-prefix.an-root.ekena' vs # f + if + ø + ekana 'nominaliser.f-reciprocal.if-0.prefix-root.ekena'; where # 0.prefix indicates the absence of a prefix; presumably the prefix an indicates # CONTROL; fifanarahana # Ungrammatical: fifarahana (0! 0 0 0) # fifanarahana f + if + an + arah + ana 'nominaliser.f-reciprocal.if-prefix.an- # stem.arah-suffix.ana' where stem arah derives from root araka 'which is able to # be followed'; vs f + if + 0 + arah + ana 'nominaliser.f-reciprocal.if-0.prefix- # stem.arah-suffix.ana'; where 0-prefix indicates the absence of a prefix; again # the prefix an indicates CONTROL; # Ungrammatical: mpifanantena (0! 0 0 0) # 'the-one(s)-who-hope...'; # from Keenan & Razafimamonjy 2001:62 cited in Hurst 2003:113, ex.136; # this is not Malagasy; mpifanantena mp+if+an+antena 'prefix.mp-reciprocal.if # -prf.an-root.antena'; in other words mp+if+anantena 'prefix.mp-reciprocal.if # -verbstem.anantena'; now the verb manantena 'to hope' is not a typical CONTROL # verb; hence the incongruity of its combination with the reciprocal if; # 6- Embedding vs Adjunction in Malagasy. # With Embedding, the matrix verb constituent-selects the tense on the lower predicate: Nikasa (ny) handeha i Paoly. # nikasa (ny) handeha i Paoly # n-i-kasa (comp) h-an-(l>d)eha D-sg Paoly # past-prf.i-root.kasa (comp) fut-prf.an-root.leha Paul # 'Paul intended to go' # The higher verb 'nikasa' constituent-selects the lower future tense, # which is obligatory;the complementizer 'ny' is optional with subject-control # predicates like 'nikasa'; note that 'ny' is the complementizer accompanying # a subject control predicate; # note that the matrix verb 'nikasa' has a tense different from the one on the # embedded verb 'handeha'. # Ungrammatical: Nikasa (ny) nandeha i Paoly. (0! 0 0 0) # nikasa (ny) nandeha i Paoly # n-i-kasa (comp) n-an-leha D-sg Paoly # past-prf.i-root.kasa (comp) past-prf.an-root.leha Paul # 'Paul intended to go'; # note that the higher verb 'nikasa' and the embedded verb 'nandeha' have # the same tense; this sequence is irretrievably ungrammatical; # With Adjunction, the main verb is accompanied by another verb with the same tense-marker: Nianatra namily fiara i Paoly. # nianatra namily fiara i Paoly. # n-i-anatra n-a(n>m)-vily fiara D-sg Paoly # past-prf.i-root.anatra past-prf.an-root.vily a/car(s) Paul #'Paul was learning how to drive a car' # Note that the matrix as well as the accompanying verbs have the same # tense; the fact that the two verbs have the same tensed-marker is in # Malagasy the hallmark of adjunction, NOT embedding; Mianatra mamily fiara i Paoly. # mianatra mamily fiara i Paoly. # m-i-anatra m-a(n>m)-(v)ily fiara D-sg Paoly # pres-prf.i-root.anatra pres-prf.an-root.vily a/car(s) Paul #'Paul is learning how to drive a car'; # note that both verbs are in the present tense; Hianatra hamily fiara i Paoly. # hianatra hamily fiara i Paoly. # h-i-anatra h-a(n>m)-(v)ily fiara D-sg Paoly # fut-prf.i-root.anatra fut-prf.i-root.vily a/car(s) Paul #'Paul will be learning how to drive a car'; # note that both verbs are in the future tense; # With Embedding, the accompanying PrepPhrase has an aspect-marker on it: Nandeha tany Antsirabe i Paoly. # N-an-(l>d)eha t-any Antsirabe i Paoly. # past-pref.an-root.leha perf-any Antsirabe D.sg Paul # "Paul went to Antsirabe." # Note that morpheme t- on the preposition 'any' is glossed as 'perfective' aspect. Tany Antsirabe no nalehan'i Jeanne. # T-any Antsirabe no n-a-leha-n(a) i Jeanne. # Perf-there Antsirabe focus past-a.passive-leha-na D.sg Jeanne # 'It was to Antsirabe Jeanne went.' # The punctual argument PP 't-any Antsirabe''perf-there' has been fronted and this is accompanied # by the A-passive voice. The DURATIVE, non-punctual passive 2 form an...an(a) is not an option, # as this would yield an irretrievably ungrammatical sequence. # This sequence shows that the PrepPhrase 'tany Antsirabe' is an argument of the motion verb 'nandeha'; # when this argument PP gets fronted, the verb must be in the circumstantial passive voice, in this case # with the a-passive form, involving PUNCTUAL INCEPTIVE aspect. # Note that the morpheme t- is ambiguous between a perfective aspect reading and a past tense reading. # With Adjunction, the accompanying PrepPhrase has the same tense-marker as the verb on it: Nijanona tany Antsirabe i Paoly. # N-i-janona t-any Antsirabe i Paoly. # past-pref.i-root.janona past-any Antsirabe D.sg Paul # "Paul stopped at Antsirabe." # Note that morpheme t- on the preposition 'any' is glossed as 'past-tense', as the verb 'nijanona' # as used here does not describe a DELIBERATE kind of ACTIVITY, but rather a STATIVE type of # situation. Tany Antsirabe no nijanona i Paoly. # T-any Antsirabe no n-i-janona i Paoly. # past-any Antsirabe focus past-pref.i-root.janona D.sg Paul # "It was at Antsirabe that Paul stopped." # Note that morpheme t- on the preposition 'any' is glossed as 'past-tense', as the verb 'nijanona' # as used here does not describe a DELIBERATE kind of ACTIVITY, but rather a STATIVE type of # situation; as such, the PP is an adjunct to the verb and when it gets fronted, the verb remains # in the active voice. # Note that the morpheme t- is ambiguous between a perfective aspect reading and a past tense reading. Hijanona any Antsirabe i Paoly. # H-i-janona ø-any Antsirabe i Paoly. # fut-pref.i-root.janona nonpast-any Antsirabe D.sg Paul # "Paul will stop at Antsirabe." # Note that morpheme ø- on the preposition 'ø-any' is glossed as 'nonpast-tense', as the verb 'hijanona' # as used here does not describe a DELIBERATE kind of ACTIVITY, but rather a STATIVE type of # situation. ø-any Antsirabe no hijanona i Paoly. # ø-any Antsirabe no ø-i-janona i Paoly. # nonpast-any Antsirabe focus nonpast-pref.i-root.janona D.sg Paul # "It is at Antsirabe that Paul will stop." # Note that morpheme ø- on the preposition 'ø-any' is glossed as 'nonpast-tense', as the verb 'hijanona' # as used here does not describe a DELIBERATE kind of ACTIVITY, but rather a STATIVE type of # situation; as such, the PP is an adjunct to the verb and when it gets fronted, the verb remains # in the active voice. # Note that the morpheme ø- is ambiguous between a nonperfective aspect reading and a nonpast tense reading. # Nominalization with prefix f on the passive 2 (circumstantial/relative) # form of the verb: Ny fikasan'i Paoly handeha # Ny fikasa-n' i Paoly handeha # Det-unspec f -i -kasa -n' i Paoly h-an-leha # Spec nprf-prf-intent-(a)n(a) D-sg Paul fut-prf-go # the intention of Paul to go; # where the verb mikasa m+i+kasa 'pres-prefix.i-root.kasa''to intend' yields # a passive 2 form ikasana i+kasa+ana 'prefix.i-stem.kasa-suffix.ana' to # which the nominalising prefix f is added; with stem kasa, stress falls on # 'sa' and NOT on 'ka', as was the case with the corresponding root. Ny fianaran'i Paoly namily fiara # Ny fianaran' i Paoly namily fiara # Det-unspec f-i-anar-an'i Paoly n-an-vily fiara # Spec nprf-prf-root-advice-by D Paul past-prf-root-drive car # the learning by Paul of driving a car; # where the verb mianatra m+i+anatra 'pres-prefix.i-root.anatra' yields a # passive 2 form ianarana i+anar+ana 'prefix.i-stem.ianar-suffix.ana' to # which the nominalising prefix f is added; with stem ianar, stress falls # on 'nar' and NOT on 'an', as was the case with the corresponding root. Ny fifanarahana # 'the mutual understanding'; # f + if + an + arahana = nominalizer prefix f + reciprocal if + lexical # causative an + affixal passive on verb stem anarahana an+arah+ana 'prefix.an- # stem.arah-suffix.ana'; anarahana is the passive 2 form for the verb manaraka # m+an+araka 'pres-prefix.an-root.araka'; again with stem anarah, stress falls # on 'rah' and NOT on 'a' of root 'araka'. # Nominalization with prefix t on the relevant phrase: # As illustrated earlier in relation to Adjunction in Malagasy, prefix t- is used very productively in # this language ambiguously either as a tense or an aspect-marker. However, in addition to this function # prefix t- can also be used in nouns derived from adjectives or prepositions although this does not seem # to be quite as productive nor as predictable: ambany # "below" PP ambanin'ny latabatra # ambani-n(a) ny latabatra # below-of the table' # "below the table"; ambany # "low" Adj ambany toerana # low location # note the preposed Adj, which normally will follow its noun; tambany # "inhabitants from lowland regions" Noun # diachronically t + ambani tambanivohitra # "inhabitants from countryside" Noun # diachronically t + ambani + vohitra # literally "from low village" ambony # "above" PP ambonin'ny latabatra # amboni-n(a) ny latabatra # above-of the table' # "above the table"; # Ungrammatical: tambony (0! 0 0 0) # "inhabitants from highland regions" Noun does not exist; tantsaha # "peasants" Noun; # diachronically t + an + t + saha # prefix.t + prep.an + epenthetic.t + saha # literally "coming from on-the fields"; tantsambo # "sailors" Noun; # diachronically t + an + t + sambo # prefix.t + prep.an + epenthetic.t + sambo ## literally "coming from on-the ship"; # 7- Particles like ho, no, etc... # Comp ho introducing a Small Clause, a non-verbal predicate describing a STATE. Milaza azy ho mahay i Paoly. # Milaza azy ho mahay i Paoly. # M-i-laza azy ho ma-hay i Paoly # comp-ho STATE-predicate # pres-say him comp intelligent D Paul # 'Paul considers himself intelligent'; note that the complementizer 'ho' # accompanies a STATE-predicate 'mahay''(be)-intelligent'; # Ungrammatical: Nifampilaza ho namboly vary Rasoa sy Ravelo.(0! 0 0 0) # N-if-amp-i-laza ho namboly vary Rasoa sy Ravelo. # pst-rec-act-say Comp pst-cultivate rice Rasoa and Ravelo # lit 'Rasoa and Ravelo said each other to have cultivated rice'; # Keenan & al. (2001: 50-51) repeated in Hurst (2003:67 ex 78); # note that in the above sequence the complementizer 'ho' accompanies a predicate # 'namboly' 'cultivated', which contrary to expectations does NOT indicate a STATE; # Ungrammatical: Nifampilaza ho namboly vary Rasoa sy Ravelo. (0! 0 0 0) # N-if-amp-i-laza ho n-amboly vary Rasoa sy Ravelo. (0! 0 0 0). # past-rec-act-prf-say comp-ho past-cultivate rice Rasoa and Ravelo. # comp-ho non-STATE-predicate # This sequence is simply not Malagasy!At best, it would be Foreigner Talk. # A major problem has to do with the lexical entry for 'nifampilaza'. The # relevant lexical frame for this complex predicate is: nifampilaza zavatra # (something-NP direct object) plural subject NP.' It thus rules out an # embedded structure with a small clause with comp-ho. # Comp ho introducing a COUNTERFACTUAL claim clause with a non-stative verb: Nilaza an-dRasoa ho namboly vary Ravelo. # N- i- laza an- d Rasoa ho n-a(n>m)-(v>b)oly vary Ravelo. # past-prefix.i-root.laza acc-ep.d-Rasoa part past-cultivate rice Ravelo # lit. 'Ravelo said Rasoa to have cultivated rice' # 'Ravelo said Rasoa cultivated rice' in Keenan & al. (2001: 50-51) as reported # in Hurst (2003:67 ex 78); # New translation: 'Ravelo falsely accused Rasoa of cultivating rice'; # note that in this case 'ho' is glossed as 'particle' and NOT as complementizer. # Charles Randriamasimanana, PhD in Linguistics (Malagasy, Austronesian)