handbücher zur sprach- und kommunikations- wissenschaft filehandbücher zur sprach- und...

Handbücher zurSprach- und Kommunikations-wissenschaftHandbooks of Linguisticsand Communication Science

Manuels de linguistique etdes sciences de communication

Mitbegründet von Gerold Ungeheuer (†)Mitherausgegeben 1985�2001 von Hugo Steger

Herausgegeben von / Edited by / Edites parHerbert Ernst Wiegand

Band 17.2

Walter de Gruyter · Berlin · New York

Bereitgestellt von | Universitäts- und Stadtbibliothek KölnAngemeldet | 134.95.162.246

Heruntergeladen am | 03.06.13 09:47

MorphologieMorphologyEin internationales Handbuch zur Flexion undWortbildungAn International Handbook on Inflection andWord-Formation

Herausgegeben von / Edited byGeert Booij · Christian Lehmann · Joachim Mugdan ·Stavros Skopeteasin collaboration with Wolfgang Kesselheim

2. Halbband / Volume 2

Walter de Gruyter · Berlin · New York



�� Gedruckt auf säurefreiem Papier, das dieUS-ANSI-Norm über Haltbarkeit erfüllt.

ISBN 3-11-017278-X

Bibliografische Information der Deutschen Bibliothek

Die Deutsche Bibliothek verzeichnet diese Publikation in der DeutschenNationalbibliografie; detaillierte bibliografische Daten sind im Internetüber �http://dnb.ddb.de� abrufbar.

” Copyright 2004 by Walter de Gruyter GmbH & Co. KG, 10785 Berlin.Dieses Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb derengen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlages unzulässig und strafbar. Dasgilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung undVerarbeitung in elektronischen Systemen.Printed in GermanySatz: META Systems GmbH, WustermarkDruck: Tutte Druckerei GmbH, SalzwegEinbandgestaltung und Schutzumschlag: Rudolf Hübler, Berlin



1473136. Tagalog (Austronesian)

Universals”. In: Derbyshire, Desmond & Pullum,Geoffrey (eds.), Handbook of Amazonian Lan-guages, Vol. I. Berlin: Mouton de Gruyter, 440�

465

Payne, Doris L. (1986 b), “Noun Classification inYagua”. In: Craig, Colette (ed.), Noun Classes andCategorization. Amsterdam: Benjamins, 113�131

Payne, Doris L. (1990 a), The Pragmatics of WordOrder: Typological Dimensions of Verb Initial Lan-guages. Berlin: Mouton de Gruyter

Payne, Doris L. (1990 b), “Morphological Charac-teristics of Lowland South American Languages”.In: Payne, Doris L. (ed.), Amazonian Linguistics:Studies in Lowland South American Languages.Austin: Univ. of Texas Press, 213�241

Payne, Doris L. (1992), “Narrative Discontinuityvs. Continuity in Yagua”. Discourse Processes 15,375�394

Payne, Doris L. & Payne, Thomas E. (1990), “Ya-gua”. In: Derybshire, Desmond & Pullum, Geof-frey (eds.), Handbook of Amazonian Languages,Vol. II. Berlin: Mouton de Gruyter, 249�474


1. Introduction2. Morphosyntax and parts of speech3. Formal processes4. Voice, aspect, and mood5. Actor involvement6. Illustrative text7. Uncommon abbreviations8. References

1. Introduction

1.1. The language and its speakers

Tagalog, a Meso-Philippine language belong-ing to the Western Malayo-Polynesianbranch of the Austronesian family, is one ofthe major languages of the Republic of thePhilippines (Constantino 1971: 112�118). Inthe 1990s, it was the native language ofaround 15 million speakers. Most Tagalogspeakers live in the central parts of the islandof Luzon, in particular in the provinces Ba-tangas, Bulacan, Laguna, and Marinduque.But Tagalog is also widely spoken in the sur-rounding provinces (including the capital,Manila), and there are many settlements of

Payne, Thomas E. (1983), “Subject Inflection ofYagua Verbs”. Ms.

Payne, Thomas E. (1993), The Twins Stories: Par-ticipant Coding in Yagua Narratives. Berkeley:Univ. of California Press

Powlison, Esther (1971), “The Suprahierarchicaland Hierarchical Structures of Yagua Phonology”.Linguistics 75, 43�73

Powlison, Esther & Powlison, Paul (1958), “El sis-tema numerico del yagua (pebano)”. Tradicion, Re-vista Peruana de Cultura 21, 3�8

Powlison, Paul (1962), “Palatalization Portman-teaus in Yagua (Peba-Yaguan)”. Word 18, 280�299

Powlison, Paul (1985), Yagua Mythology: EpicTendencies in a New World Mythology. Dallas: In-ternational Museum of Cultures

Rivet, Paul (1911), “La famille linguistique Peba”.Journal de la Societe des Americanistes de Paris 8,173�206

Doris L. Payne, Eugene (U.S.A.)

Tagalog speakers throughout the Philippineislands, especially on Mindoro and Minda-nao (cf. McFarland 21983: 29, 80). Further-more, Tagalog is widely spoken as a secondlanguage. In 1937 it was selected as the basisfor the national language Filipino (formerlycalled Pilipino).

Since the beginning of Spanish colonisa-tion in 1565 the Tagalog area has always beenthe center of political power in the Philip-pines, and Tagalog has thus been strongly in-fluenced by the colonial languages, firstSpanish and then American English (from1898 to 1946). This influence, however, hasbeen largely on the lexicon and the phonol-ogy, but not on the morphosyntax. For ex-ample, the Manila slang, called Taglish,mainly consists of English roots, but themorphology is exclusively Tagalog. Examplesinclude lipstick-an ‘to put lipstick on’, i-giveup ‘to give sth. up’, and mag-on ‘to start dat-ing’ (cf. Cubar 1984; the affixes are explainedin 4.1). As for morphosyntax, it is not un-likely that there has been some normative in-fluence by colonial institutions and practices



1474 XVI. Systeme morphologischer Struktur: Sprachskizzen

because the Spanish started using Tagalogvery early on as a missionary language, writ-ing grammars and preparing catechisms inTagalog (a Doctrina Christiana, dated 1593,is the oldest preserved Tagalog document).The continuing normative influence exertedby the Manila based educated classes isshown by the fact that Tagalog exhibits sur-prisingly little dialectal diversity (McFarland21983: 80). Only the dialect spoken on thesmall island of Marinduqe exhibits lexicaland morphological idiosyncrasies that havecaught the attention of linguists (Lopez 1970;Soberano 1980).

The most influential of the Spanish gram-mars is the one written by Totanes (41865)which has served as the basis for many analy-ses of Tagalog (e.g. Humboldt 1838; Müller1882: 87�163; Marre 1901). The last andmost comprehensive Tagalog grammar castin a traditional framework is Blake (1925).Bloomfield’s texts and grammar (1917) arethe first attempt to present Tagalog in its ownterms and a very fine and early piece of mod-ern structuralist analysis. In addition toBloomfield and Blake, further comprehensivestatements of Tagalog morphology can befound in Lopez (1937), Schachter & Otanes(1972), and Wolff et al. (1991). A review oflinguistic work on Tagalog is given in Con-stantino (1971: 118�145) and Reid (1981).The present article is based primarily onBloomfield’s data and analysis.

1.2. Phonology and orthography

The segmental phonology of Tagalog is rela-tively simple (except for problems concerningthe treatment of loans for which see MatsudaFrench 1988:1�17). The graphemes used inthe standard orthography directly reflectphonemes and basically have their IPA val-ues. The only exception is the use of the di-graph <ng> for the velar nasal [n]. The useof this digraph may be slightly confusing,since it is also used for the NP-marker [nan],in this case written as a separate word whichin turn is not to be confused with -ng � [n],the bound allomorph of the linker na. A gen-uine problem with the standard orthographyis the fact that glottal stops are not written.Glottal stops regularly occur before initialvowels and intervocalically, i.e., <tao> ‘per-son’ is [ta?o], <aral> ‘study’ is [?aral] and<magaaral> ‘will study’ is [mag?a?aral]. Asthe last example shows, base-initial glottals

are not dropped when prefixes are added. Al-though the phonemic status of these initialand intervocalic glottals is far from clearmuch recent writing on infixation in Tagalogassumes without any discussion that they arenon-phonemic (e.g. Crowhurst 1998: 590 whomisrepresents her older sources with regardto this point). Word-finally, the problem isfurther confounded by the fact that here finalvowels (optionally followed by [h]) contrastwith glottal stops. Thus, <baga> ‘glowingcharcoal’ is [bagah], but <baga> ‘lungs’ is[baga?]. Deviating from standard orthogra-phy a final glottal stop will be marked in thisarticle by <’>.

Stress is also unmarked in standard or-thography. Its analysis is somewhat contro-versial. Some authors (e.g. Schachter &Otanes 1972: 15�18; Wolff et al. 1991: 12)consider vowel length the primary phenome-non, while others consider vowel length anepiphenomenon of stress (cf. Bloomfield1917: 141 f.; Matsuda French 1988: 63 f.). Inthis article, the latter view has been adopted,but since stress assignment in Tagalog is notfully understood, all remarks pertaining tothis phenomenon must be regarded with cau-tion. Stress clearly is phonemic in Tagalog,compare bukas ‘tomorrow’ and bukas ‘open’,and plays an important role in affixation (see3.4). Bases with stress on the penultimate syl-lable are called barytone bases, those withstress on the final syllable oxytone bases. Pri-mary stress on the penultimate syllable willremain unmarked (thus bukas for ‘tomor-row’), elsewhere it will be marked by theacute accent. The grave accent marks second-ary stress.

Phonological rules of relevance to thepresent exposition are:

� /d/ often becomes /r/ intervocalically;� phrase-final vowels are followed by a

weak glottal fricative (cf. [bagah] above);� /u/ and /i/ in phrase-final syllables are

lowered to [o] and [e], respectively;� word-final glottal stops regularly disap-

pear before clitics and the linker (cf.Bloomfield 1917: 136; Wolff et al. 1991:12).

2. Morphosyntax and parts of speech

Tagalog distinguishes between two parts ofspeech (Bloomfield 1917: 146): full words andfunction words (or particles). Some function




words mark morphosyntactic slots, then usu-ally being proclitics. Others are second posi-tion clitics (these are not discussed here; seeSchachter & Otanes 1972: 187�193, 433�435and Kroeger 1993: 118�123, 152�154). Fullwords may be extensively affixed and occurin the limited set of morphosyntactic slotsprovided for by the function words. Thereare five morphosyntactic slots for full wordsin a Tagalog clause, four of which are il-lustrated by the following example:

(1) i-ni-abotcv-rls(ugr)-within.reach

nggen

mang-ga-gamotav-rdp2-medicine

saloc

sundalosoldier

angspec

itlog,egg

atand

angspec

parepriest

atand

siya3.sg

aypm

nag-hintayrls.av-wait

nggen

sa-sabih-inrdp1-say-pv

nggen

sundalo.soldier

‘The physician handed the egg to the sol-dier, and the priest and he waited forwhat the soldier would say.’

Except for the first word, all full words inthis clause are preceded by a function word.The clause initial position is the predicate po-sition, which is unmarked unless it is pre-ceded by a topical constituent. In this case,the predicate is marked by the predicatemarker ay ‘pm’, as in the second part of ex-ample (1). The other markers have the fol-lowing functions:

sa is a general locative preposition mark-ing oblique arguments and adjuncts. It is thefinal constituent of all of the more specificprepositions in Tagalog such as hanggang sa‘until’ or para sa ‘for’.

ang is often called a topic (or a subject)marker in the literature. The notion of sub-ject in Tagalog is highly controversial (cf.Schachter 1976; 1995; Drossard 1984: 73�78;Foley & van Valin 1984: 134�148; DeWolf1988: 144�150; Kroeger 1993; Naylor 1995).The fact that ang by itself marks neither top-ics nor subjects is evident from clauses con-taining two ang-phrases:

(2) angspec

mgapl

buhokhair

lamangonly

angspec

p<in>u-putol<rls(ugr)>rdp1-cut

nggen

patalımblade

‘only the hair was cut by the blade’

In this example, the first ang-phrase func-tions as the predicate, the second as the pred-ication base or subject (subject2 in the termi-nology of Matthews 1981: 104�113). Sub-jects in this sense have to be configurationally

defined in Tagalog as the ang-phrase whichoccurs either after the predicate or before thepredicate marker ay. But this does not yetclarify the function of ang.

Basically, the function of ang is similar tothat of an article (the standard analysis intraditional Tagalog grammars, cf. Blake1925: 205 f.). It is, however, not a definite ar-ticle, but includes all kinds of referentiallyspecific expressions (definite, specific-indefi-nite, generic; for details, see Adams & Man-aster-Ramer 1988 and Himmelmann 1991:8�16), and is therefore glossed as spec here.

ng [nan] ‘gen’ marks genitive attributes. Inthe literature it is common to differentiate be-tween ng marking non-topic agents, directobjects, instruments, manner, etc., but thereis little empirical support for such distinc-tions (cf. Naylor 1980: 37�42).

The fifth morphosyntactic slot, not il-lustrated by examples (1) and (2), is consti-tuted by the linker (or ligature) na (-ng aftervowels, /n/ or glottal stop). This particle linksthe elements of a modifying constructionsuch as ulol na unggo’ ‘foolish conn monkey’,but also occurs in compounds such as puno-ng-saging ‘tree-conn-banana’ and in comple-ment clauses (see Gonzales 1971). The orderin modifying constructions is not fixed in Ta-galog, thus unggo-ng ulol is equally possiblefor ‘foolish monkey’. The difference betweenthe genitive marker ng and the linker pertainsto referentiality as shown by the following‘minimal pair’:

(3) (a) bata-ngchild-conn

dalagayoung.woman

‘girl’(b) bata’

childnggen

dalagayoung.woman

‘child of the young woman’

Full words are not formally differentiatedwith respect to the five morphosyntactic posi-tions just illustrated. Any full word, regard-less of its affixation, may occur in any posi-tion (provided its meaning is appropriate).That is, with regard to their syntactic distri-bution the large class of full words cannotbe further divided into classes such as nouns,verbs, and adjectives (cf. Lemarechal 1982;1989; Himmelmann 1991; to appear; Gil1993; Shkarban 1995; Naylor 1995; Art. 72).For example, the words sasabihin in (1) andpinuputol in (2) which are inflected for aspectand mood (see 4.2) and which are oftencalled verbs, occur � without any further




derivation � in a ng- and an ang-phrase,respectively. Similarly, so-called nouns mayoccur underived in predicate position, e.g.

(4) Isda’fish

angspec

pag-kainger-eat

niya.3.sg.poss

‘His meal was fish.’

A brief look at the list of affixes inSchachter & Otanes for what they call‘nouns’ (1972: 97�106), ‘adjectives’ (1972:198 f.; 216�229), and ‘verbs’ (1972: 344�355)immediately reveals that a subclassificationof full word bases on the basis of morpho-logical evidence is also not a straightforwardenterprise since basically the same set of af-fixes is involved in all these formations. Nev-ertheless, bases appear to differ in their mor-phological potential. A detailed study of thedifferent morphological classes, however, stillremains to be done (see Wolff 1993 and Him-melmann to appear for some preliminarysuggestions).

There is no doubt that both unaffixedbases and affixed words differ in their seman-tics in that some denote actions, othersthings, yet others states, etc. A base such asbilı clearly designates the action of buying,bahay the entity ‘house’, bago the quality‘new’. The distinction between these differentkinds of concepts is sometimes grammaticallymarked by different stress patterns (cf. 3.4).It is not clear, however, whether these con-ceptual classes are in any way directly rel-evant to stating the (segmental) morphosyn-tactic regularities of Tagalog. Therefore,given the absence of clear-cut formal evi-dence for such categories, the terms for lexi-cal categories such as noun, verb, adjective,etc. are avoided in this article and the termsaction, entity, etc. are used whenever it isconvenient to make reference to conceptualclasses.

3. Formal processes

3.1. AffixationThe number of affixes in Tagalog is fairlysmall, but each of the major three affix cat-egories is represented: there are two suffixes(-in and -an, see 4.1), two infixes (-in- and-um-, see 4.1 and 4.2), and about a dozen pre-fixes. Most of these affixes can be combinedwith each other so that a large number ofcomplex formations results (Bloomfield1917: 317�319 lists some 200 formations).This number is further increased by the fact

that affixes may co-occur with other formalprocesses such as reduplication (see 3.3) andstress shifting (see 3.4).

Prefixes exhibit some word-like propertiesin that (with few exceptions) they (a) do notdisplay any fusional characteristics (not evenresyllabification takes place, which, however,is not surprising given the fact that all basesstart with a consonant (cf. 1.2)), (b) mostlycarry their own stress, and (c) may occasion-ally be used in isolation (cf. Bloomfield1917: 213; Matsuda French 1988: 89 f.; Ru-bino 1998). Exceptions are, on the one hand,the prefixes i- (4.1) and ka- (5.1), the first ofwhich is never stressed and both of which aresometimes fused with the base, e.g. ka-ibig-an ‘ass-like-lv (friend)’ is usually pronounced[kaybigan] rather than [ka?ibigan] (Bloom-field 1917: 139 f.).

The other exception is the prefix pang-(and related mang- and nang-), where the pre-fix-final nasal regularly assimilates to, andsometimes substitutes for, the initial conso-nant of the base. This prefix-type is verycommon in western Austronesian languagesand is often analysed as containing an archi-segmental nasal (and then is represented as/paN-/). The assimilation and substitutionregularities may be summarised as follows(for the semantics of this derivation, see 5.2):

� N is regularly deleted before base-initialnasals: maN � manhıd J mamanhıd ‘getnumb’.

� N never substitutes for a glottal fricativeor for glides: maN � hiya’ J manghiya’‘humiliate’, maN � walıs J mangwalıs‘hit with a broom’, maN � yari Jmangyari ‘happen’.

� Voiceless obstruents (apart from the glot-tal fricative) are regularly substituted for:maN � pili ’ J mamili ’ ‘choose (severalthings)’, maN � takot J manakot‘frighten several people’, maN � kabayoJ mangabayo ‘ride on horseback’, maN� sakıt J manakıt ‘cause pain’. Glottalstop is sometimes not substituted for; thefactors involved are unclear. In fact, twoderivations are occasionally possiblefrom the same base. Compare [manaso]‘hunt with dogs’ with [man?aso] ‘ridiculepeople’ (base aso ‘dog’).

� Voiced obstruents and the lateral are gen-erally not substituted for: maN � daya’J mandaya’ ‘cheat people’ (but maN �dikıt J manikıt ‘adhere’), maN � gamotJ manggamot ‘practice medicine (profes-




sionally)’ maN � laro’ J manlaro’ ‘amuseothers’. For /b/ both options are commonand, in a few cases, possible for the samebase. Thus from bakya’ ‘wooden shoe’ itis possible to derive either mambakya’‘hit with a wooden shoe’ or mamakya’‘wear wooden shoes’.

In the derivation of so-called instrumentaladjectives substitution is said to be optional,e.g. both pansuklay and panuklay mean ‘foruse in combing’ (Schachter & Otanes 1972:218�221; 319 f.).

Suffixes exhibit regular fusional tendenciesin that the stem-final consonant (includingthe optional [h] after vowel-final bases) isregularly resyllabified with the suffix and thebase-final vowel is often syncopated, e.g. la-kas � an J lak.san ‘strengthen (x)’, bilı � inJ bil.hın ‘buy (x)’ (cf. Matsuda French 1988:12 f.). There are further sporadic irregulari-ties such as tawan-an ‘laugh at (x)’ instead ofthe expected *tawah-an, and tamn-an ‘plantin (x)’ instead of the expected ?tanim-anor *tanm-an (cf. Bloomfield 1917: 214;Schachter & Otanes 1972: 375�380).

Infixes are always inserted after the firstconsonant of the base, e.g. b<um>ilı ‘<av>buy’ from bilı. All other formative proces-ses precede infixation, infixes thus alsooccur in prefixes, e.g. p<in>ag-tapat-an‘<rls(ugr)>ger-fronting-lv (confided to(x))’, and reduplicated syllables (cf. 3.3). Informations involving the infix -in- and stem-initial glides or liquids, the stem-initial con-sonant and the nasal of the infix are regularlymetathesized, thus *l-in-uto’ but ni-luto’‘rls(ugr)-cook’. With stem-initial /w/ me-tathesis is optional, for example, both ni-wisik-an and w<in>isik-an for ‘rls(ugr)-sprinkle-lv/<rls(ugr)>sprinkle-lv (sprinkleon (x))’. When co-occurring with the prefixi- (i.e. [?i]), -in- is infixed into the followingsyllable, e.g. i-b<in>igay ‘cv-<rls(ugr)>give’ from bigay. In stems beginning with aliquid, a glide or a glottal, metathesis againregularly occurs, compare i-ni-hatıd ‘cv-rls(ugr)-convey’ (*ihinatid), i-ni-abot [?i-ni-?abot] ‘cv-rls (ugr)-within.reach’ (*i-in-abot).

3.2. Consonant alternation

A number of prefixes display regular alterna-tion of the initial consonant (cf. de Guzman1978: chapter 3.3): the /p/-initial form is thebasic form (used as gerund (4.3) or impera-tive), /m/ marks non-realis and /n/ realis (cf.

4.2). Examples are pag-/mag-/nag-, paN-/maN-/naN-, paki-/maki-/naki-. The alterna-tion between the two nasal forms also occurswith the stative or potentive prefix (ma-/na-,cf. 5.1), but here no /p/-initial basic form ex-ists. Furthermore, a small number of /p/-ini-tial bases (which probably contain a fossilisedprefix) exhibits this alternation, e.g. pakinıgmakinıg nakinıg ‘listen’, panood manood na-nood ‘watch’. This alternation probably de-veloped from infixed formations by clippingthe first (unstressed) syllable. Thus, mag-probably derives from *pumag- and nag-from *pinag-. Note, however, that pinag- isstill a productive formative, i.e. the rls(ugr) form of pag-derived stems (cf. 5.2).

3.3. ReduplicationThere are three kinds of reduplication pro-cesses in Tagalog. In two of these processesthe first consonant and vowel of the base arecopied (CV-reduplication). The two processesdiffer with regard to the fact that in one thereduplicated syllable is assigned stress (rdp1),and in the other it is not (rdp2): mang-ga-gamot ‘will practice medicine’ vs. mang-ga-gamot (or manggagamot) ‘one who makescures, physician’. If a base starts with a con-sonant cluster, only the first consonant iscopied, cf. trabaho ‘work’ J mag-ta-trabaho‘will work’.

The third process consists in copying thefirst two syllables of the base (rdp3). In thecase of disyllabic bases, the complete base iscopied, e.g. lakad ‘walk’ J mag-lakad-lakad(rdp3 with stress shift) ‘do a little walking’,mura ‘cheap’ J mura-mura ‘rather cheap’. Inthe case of bases containing more than twosyllables, the second syllable is copied only upto its peak, which is then assigned secondarystress, viz. tahimik ‘peaceful’ J tahı-tahimik‘rather peaceful’, baluktod ‘crooked’ J balu-baluktod ‘variously bent’. rdp2 and rdp3 gen-erally apply to bases only, while rdp1 freelyapplies to prefixes as well (see below).

Each type of reduplication may occur onlyonce in a derivation but different types maybe combined with each other, as will be seenshortly.

Tagalog reduplication phenomena havefigured prominently in the discussion of thestatus of reduplication (cf. Art. 57) as well asthe related issues concerning the boundariesbetween morphology and phonology and theinternal structure of the morphological com-ponent (cf. Marantz 1982: 438 f., 451 f., 473�479; Carrier-Duncan 1984; Matsuda French




1988: 19�61). The argument is concernedwith the interrelation of (regular) affixationand reduplication. The Tagalog evidence,however, seems inconclusive in this regard.The following is a (partial) list of the rel-evant phenomena:

� Formations involving nasal substitution(cf. 3.1) suggest that allomorphy rulesprecede reduplication, e.g. maN � rdp1� putol J mamumutol ‘will cut (a lot)’.

� Infixation, on the other hand, seems tofollow reduplication (cf. Matsuda French1988: 38), i.e., pumuputol ‘cut’ seems tobe derived from puputol, otherwise onewould expect *pupumutol.

� In formations involving polysyllabic pre-fixes or a combination of prefixes, it isusually the second syllable of the prefixthat is reduplicated (by rdp1, receivingword level stress, cf. Matsuda French1988: 44�52), viz. mag � pa � putol Jmagpapaputol ‘will cause to be cut’, maka� putol J makakaputol ‘will be able tocut’, or maka � pag � pa � putol J ma-kakapagpaputol ‘will be able to cause tobe cut’. A general exception to this regu-larity are formations involving the prefixi-, which is never reduplicated and doesnot count in determining the second syl-lable, cf. ma � i � pag � luto’ J maipa-pagluto’ ‘will be able to cook for (x)’ andi � paki � pa � putol J ipakıkipaputol‘will be asked to be caused to be cut’. Inseveral instances, however, either the sec-ond syllable of the prefix or the first sylla-ble of the base may be reduplicated, anexample being makapuputol ‘will be ableto cut’ which alternates with the formmakakaputol just mentioned (for moreexamples, cf. chart in Schachter & Otanes1972: 369). This raises the problem ofwhere to place rdp1-reduplication in thederivation of these complex formations.

� The relation of suffixation and rdp3 isalso problematic. In some cases suffix-ation (and the related morphonologicalprocesses) clearly apply before rdp3:tingın � an J tingnan ‘look at (x)’ J ting-nan-tingnan ‘look at (x) a little’. In othercases, it is the other way around: sakit Jsakit-sakit J magsakıt-sakitan ‘pretendto be sick’. If suffixation preceded rdp3,*magsakisakitan would be expected (tri-syllabic base), cf. mag � salita’ � an Jmagsalı-salitaan ‘talk a little to eachother’.

� rdp1 and rdp3 may co-occur. In this caseeither rdp3 precedes rdp1, e.g. magsa-sakıt-sakitan ‘will pretend to be sick’ andmagsasalı-salitaan ‘will talk a little toeach other’, or they may apply simulta-neously (at different locations), viz. mag� pa � ka � ingat J magpapakaingat-ingat ‘will be extremely careful’. Carrier-Duncan (1984: 269) claims that there arealso cases where rdp2 precedes rdp3: mag� kaN � punit J magkampupunit ‘willtear spontaneously’ J magkampupu-pu-punit ‘will (intensively) tear spontane-ously’.

� rdp1 and rdp2 may co-occur as well, inwhich case rdp2 precedes rdp1. Thusfrom takbo ‘run’ mag-ta-takbo ‘run wild’is derived by prefixing mag- and rdp2.From the latter the imperfective aspectmag-ta-ta-takbo is derived by rdp1.

3.4. Stress shiftBoth primary and secondary stress assign-ment in Tagalog may be connected with avariation in meaning and this assignment isat least partially independent of segmentalprocesses (cf., for example, the difference be-tween rdp1 and rdp2 mentioned in 3.3). Thedata, however, are not clear, since mostsources (apart from Bloomfield 1917 andWolff et al. 1991) do not mark stress consis-tently (cf. 1.2). Two examples are given to il-lustrate the possibilities and complexities in-volved.

There is a substantial number of baseswhich differ only with respect to stress andwhich are clearly semantically related (unlikethe pair bukas ‘tomorrow’/bukas ‘open’ men-tioned in 1.2). Examples include abot ‘over-take’ vs. abot ‘within reach, reach for, pass’,alam ‘knowledge’ vs. alam ‘known’, buhay‘live, life’ vs. buhay ‘alive’, bunot ‘pull out’ vs.bunot ‘pull out a lot/repeatedly’, isip ‘think,thought’ vs. isıp ‘think hard/with delibera-tion’, lakad ‘walk’ vs. lakad ‘on foot, bare-footed’, tulog ‘sleep’ vs. tulog ‘asleep’ (cf.Bloomfield 1917: 215 f.). Wolfenden (1961:12) characterises the meaning shifts involvedas (a) accomplishment (resultative) or (b) in-tensification. While resultative pairs arewidely attested, it is unclear whether themuch more sporadic pairs not belonging tothis type can all be subsumed under ‘intensifi-cation’ as the handful examples just givenshould make clear.

The complex interaction of stress assign-ment and affixation is illustrated by the suffix




-an (cf. Bloomfield 1917: 250�262). If thissuffix marks locative voice (cf. 4.1), primarystress usually shifts one syllable to the right(i.e. to the ultimate or penultimate syllable ofthe derived word). Examples are tawag ‘call’J tawagan ‘call (x)’, gupıt ‘cut (with scissors)’J gupitan ‘cut the hair of (x)’, bilı ‘buy’J bilhan ‘buy from (x)’, sama ‘go along, ac-company’ J samahan ‘accompany (x)’. If -anderives expressions denoting either a collec-tive action or the place where something (anentity or an action) is located, stress in oxy-tone bases remains on the same syllable as inthe underived word: iyak ‘cry, weep’ J iya-kan ‘a crying of many’, ‘buy’ J bilıhan ‘placewhere to buy, market’, aklat ‘book’ J akla-tan ‘library’, litson ‘a roast pig’ J litsunan‘place to roast pigs or a barbecue (� roast-pig party)’. For barytone bases there are twopossibilities: either stress shifts one syllable tothe right (to the penultimate syllable of thederived word), which is often accompaniedby secondary stress on the first syllable of thederived word, e.g. samahan ‘a going togetherof many, company’, or it is shifted to the ulti-mate syllable of the derived word (i.e., thesuffix is stressed), e.g. basa ‘read’ J basahan‘a reading-room, library’ (vs. basahan ‘readsth. to (x)’), laban ‘contrary, fight’ J labanan‘a fighting of many, battle, war’ (vs. labanan‘fight/oppose (x)’). Stress shift to the penulti-mate syllable of the derived word is used(with very few exceptions) for collective ac-tion expressions, while stress shift to the suf-fix is more commonly (though by no meansexclusively) used for ‘place where’ deriva-tions. Note that these are only regularities;differences in meaning are not always accom-panied by formal differences, e.g. dulo ‘end’J duluhan ‘terminate (x)’ (locative voice), butalso ‘end part, back yard’ (‘place where’; cf.Bloomfield (1917: 261) who lists a number ofwords which formally appear to be locativevoice derivations, but the meaning of whichdoes not fit this categorisation).

4. Voice, aspect, and mood

Tagalog and the other Philippine languagesare most famous for a phenomenon variouslycalled voice, orientation, case marking on theverb (cf. Blake 1906; Ramos 1974), or ‘focus’,a term introduced in the late 1950s to un-derscore the exceptional nature of the phe-nomenon (cf. Llamzon 1973: 168). The lastterm is widely used to refer to the pragmatic

phenomenon of highlighting new or con-trastive information. ‘Focus’-affixes in Philip-pine languages do not have such a highlight-ing function. The participant ‘focussed’ on bythese affixes is usually given information andoften remains unexpressed. Therefore, thisterm is avoided here. Instead, voice is chosenbecause it is the least misleading term (see4.1).

The literature on voice in Tagalog is fairlyextensive (see � in addition to the generalreference works mentioned in 1.1 � Müller1882: 136�142; Marre 1901: 574�582; Blake1906; Scheerer 1924; Capell 1964; Llamzon1973; 1976: 89; Wolfenden 1961: 14�16;Ramos 1971: 21�23, 56�69; 1974: 19�40;Foley 1976: 105�113; McFarland 1976: 16�24; de Guzman 1978: chapter 3; Drossard1984: 34�51; Himmelmann 1987: 92�125;DeWolf 1988; Shibatani 1988). Since voicemarking is formally tied to aspectual andmodal distinctions, these three categories willbe treated together in one section.

4.1. VoiceTagalog predicate expressions usually displaya voice affix that indicates the semantic roleof one of the participants involved in thestate of affairs denoted by the predicate.There are four such affixes, as illustrated bythe following examples:

(5) t<um>ango’<av>nod

angspec

unggo’monkey

‘the monkey nodded in assent’

(6) dikdik-ıncrush-pv

siya3.sg

saloc

lusongmortar

‘(that) he (i.e., the turtle) be crushed ina mortar’

(7) hulug-androp-lv

mo2.sg.poss

ako!1.sg

‘drop me (some)!’

(8) kungif

i-ta-tanimcv-rdp1-plant

niya3.sg.poss

angspec

kaniya-ngdat.3.sg-conn

ka-parteass-part

‘if he would plant his part (for him)’

As briefly shown in 2, the NP-markers angand ng in Tagalog do not signal semanticroles. Rather, the voice affixes indicate the se-mantic role of the participant which appearsin the ang-phrase (siya in (6) and ako in (7)are ang-forms of the pronoun). Thus, the in-fix -um- in (5) indicates that it is the monkeywho does the nodding, and in (6) the suffix




-in indicates that the turtle is going to be theundergoer of the crushing (rather than theactor), etc.

Before discussing some of the more re-markable features of this voice marking sys-tem in more detail, it should be noted thatthe actor voice marking infix -um- does notonly occur in clauses with a subject which isin full control of an action. It is also used forsubjects which are involved in a process, asin p<um>ula ‘become red’ or l<um>u-lu-tang ‘be floating’. Furthermore, it occurs insubject-less expressions for natural eventssuch as um-ulan ‘rain’ or l<um>indol ‘earth-quake’.

In addition to -um-, the prefix mag- (realisnag-) also marks actor voice, cf. nag-la-laro’sila ‘rls.av-rdp1-play 3.pl (they are play-ing)’. Following de Guzman (1978: chapter3), this prefix is analysed here as involvingthe prefix pag- used in gerund formation (see4.3), actor voice (and mood) being signalledby consonant alternation (cf. 3.2). The differ-ence between the two actor voice affixes isfurther commented upon in 5.2.

One of the remarkable features of voicemarking in Tagalog, which sets it apart fromvoice marking systems in many other lan-guages, is the fact that both actor as well asundergoer voices involve overt morphologicalmarking, while in languages such as Englishonly undergoer orientation (passive) is explic-itly marked. In other words, actor voice andundergoer voice are equally marked in Taga-log (at least in morphological terms).

Another peculiarity is the fact that there isnot only one affix for undergoer orientation.Instead, three different ways in which the un-dergoer may be involved in a given state ofaffairs are distinguished:

� -in (patient voice) indicates a directly af-fected undergoer, such as the turtle in (6),the hair in (2), or ito in inum-ın mo ito‘drink-pv 2.sg.poss prox (drink this)’.

� -an (locative voice) is used for recipients(see (7)), addressees, beneficiaries, andthe location where an action takes place,e.g.:

(9) ni-lakar-anrls(ugr)-walk-lv

ko1.sg.poss

angspec

ma-bato-ngstat-stone-conn

kalyestreet

‘I walked on a stony road’

More generally, it is used for indirect un-dergoers, i.e. undergoers which are not

directly affected by the action denoted bythe predicate, as in inum-an mo ito ‘drink-lv 2.sg.poss prox (drink from/some ofthis)’ or buks-an mo ang pinto’ ‘open-lv2.sg.poss spec door (open the door)’.

� i- (conveyance voice) indicates an un-dergoer that is moved (a displacedtheme), such as the egg in (1) or one halfof the banana tree in (8). It is also usedfor the instrumental role, instrumentsthus being conceived of as moving un-dergoers:

(10) Angspec

itakbolo

aypm

i-p<in>utolcv-<rls(ugr)>cut

ko1.sg.poss

nggen

saging.banana

‘I cut bananas with the bolo.’

Furthermore, i- may also indicate thebeneficiary of an action with a few bases(e.g. i-bilı ‘buy for (x)’), a usage not easilyaccounted for by any of the analyses pro-posed for this prefix (cf. Himmelmann1987: 103�22, 139 f.).

A third cross-linguistically remarkable fea-ture of the voice affixes is that they may beapplied to all kinds of bases without any fur-ther derivation. That is, the above affixescannot only be attached to bases denoting ac-tions but also to ones denoting things (e.g.bato ‘stone’ J batuh-ın ‘throw stones at (x)’),masses (e.g. tubig ‘water’ J tubig-an ‘addwater to (x)’), states (e.g. bago ‘new’ Jbaguh-in ‘change (x)’ or i-bago ‘move (x) toanother position’), or animate beings (e.g.langgam ‘ant’ J langgam-ın ‘be infested withants’). Of course, the derivational possibil-ities depend on the semantic compatibility ofbase and affix, and thus are more restrictedwith regard to, for example, expressions forhuman beings than for action expressions.

It is common to treat voice with respect toaction expressions as inflection, and voicewith regard to non-action expressions as deri-vation. There is, however, no formal evidenceto support this distinction. On the contrary,the analysis of voice as inflection leads to ex-tremely complex systems of ‘verb’ classes inTagalog. That there is little clear-cut evidencefor such classifications is shown by the factthat the proposed classifications differ ex-tremely. Blake (1925), for example, proposes17 classes, de Guzman (1978) about 80 (cf.Himmelmann 1987: 69, 129�145). The mainempirical observation here is that there is nosimple classification for action expressions




with regard to their voice marking. In par-ticular, there are no productive inflectionalparadigms for voice, as suggested by thecommonly used ‘paradigmatic’ examples inthe literature. Instead, derivations from allkinds bases are only partially predictable onthe basis of their semantics and exhibit alarge number of idiosyncrasies, which againsuggests derivation rather than inflection. SeeMcFarland (1976) and Ramos & Bautista(1986) for instructive surveys of those deriva-tions which are actually attested for a givenaction base (see also Art. 38).

With regard to the three features just men-tioned, Tagalog voice marking has much incommon with nominalising morphology inother languages. Like much of the morphol-ogy used for deriving nouns (or, in many lan-guages, participles) from verbs, the voice af-fixes change the orientation of a given basein such a way that it may be used to refer toone of the participants involved in the stateof affairs denoted by the base (cf. Lehmann1984: 151 f., who introduces the term orienta-tion for analysing nominalisation strategies).In this view, -um- is an actor orienting infixwhich derives from a base such as tango’‘nod, nodding in assent’ a word tumango’which could be glossed as ‘one who nods,nodder’. This expression no longer directlydenotes the action of nodding, but rather theparticipant who nods. That is, in the Tagalogclause (5) tumango’ ang unggo’ both tumango’and unggo’ refer to the same entity. Imitatingthe equational structure of this clause it couldbe rendered as ‘nodd-er in assent (was) themonkey’. Similarly, example (6) is ‘he (be)crush-ee in the mortar’, example (7) ‘I (be)the place of your dropping/your droppery’,and example (8) ‘if his plant-ee (would be) hispart’ (cf. DeWolf 1988: 157 f.). Note, how-ever, that Tagalog voice affixes are not nomi-nalising in a morphosyntactic sense, sincethey do not change the syntactic category ofthe base (both base and derived word are fullwords which can be used in any of the fivemajor morphosyntactic slots discussed in 2).That is, the similarity of Tagalog voice affixesand noun or participle-forming morphologyis primarily a semantic one.

With regard to this semantic similarity, itshould be noted that the voice marking for-matives do not only occur in expressionswhich denote actions, processes or states (or,more precisely, a participant involvedtherein). Most of the voice marking affixesalso occur in formations which clearly denote

entities. The major possibilities pertaining tolocative -an are mentioned in 3.4. Actor voicemarking mag- (plus rdp2) occurs in forma-tions denoting professionals, thus from na-kaw ‘steal’ mag-na-nakaw ‘thief’ may be de-rived. This form differs only with respect tostress from the action denoting formationmag-na-nakaw ‘will steal’ (cf. 3.3 and Bloom-field 1917: 242 f.; Schachter & Otanes 1972:103). Furthermore, mag- is used togetherwith kinship terms to indicate two (with rdp2several) persons between which the relationdesignated by the base holds: mag-ina‘mother and child’ (< ina ‘mother’), mag-pi-pinsan ‘several cousins’ (cf. Bloomfield 1917:242; Schachter & Otanes 1972: 102). The suf-fix -in may denote entities undergoing the ac-tion denoted by the base, e.g. aral ‘study’ Jaralın ‘lesson’. Again, it is stress that (often)differentiates action and thing, compare ar-alın with aralin ‘study (x)’. Other examplesare kumpunı ‘repair’ J kumpunihin ‘things torepair’ vs. kumpunihın ‘repair (x)’; kain ‘eat’J kanin ‘boiled rice’ or ‘eat (x)’ (no differencewhatsoever, compare also kakanın ‘sweets’);inom ‘drink’ J inumın ‘drinking water’ or‘drink (x)’, cf. also ınumin ‘beverage’ (cf.Bloomfield 1917: 247; Schachter & Otanes1972: 99 f.). There are no derivations with-um- or i- which denote entities in a similarway.

4.2. Aspect and moodVoice marked forms occur in two aspects(perfective and imperfective) and two moods(realis and non-realis). Imperfective aspect isindicated by rdp1 (perfective aspect is un-marked), realis mood is indicated by the infix-in- or consonant alternation (/m/ J /n/, cf.3.2), non-realis again being unmarked. Theseformations are illustrated in Tab. 136.1 withthe paradigms for the base bilı ‘buy’, one ofthe few bases which occurs with all voice af-fixes. A number of different analyses and ter-minologies have been proposed for theseforms (see Werlen in Bader et al. 1994: 95�100 and Kroeger 1993: 15�18 for overviews).The terminology used here reflects the formalmake-up of the paradigms but it is not quitefelicitous in all instances with regard to thefunction of the forms.

Aspect-mood formation is highly regularboth formally and semantically and thusclearly is an instance of inflection. It existsfor every voice-marked form and is alsofound for other affix combinations, e.g.maki-, makıki-, naki-, nakıki- (see 5.2). With




av (‘buy’) av (‘sell’) pv lv cv

non.rls/pfv b<um>ilı mag-bilı bilh-ın bilh-an i-bilınon.rls/ipfv bıbilı mag-bıbilı bıbilh-ın bıbilh-an i-bıbilırls/pfv b<um>ilı nag-bilı b-in-ilı b-in-ilh-an i-b-in-ilırls/ipfv b<um>ıbilı nag-bıbilı b-in-ıbilı b-in-ıbilh-an i-b-in-ıbilı

Tab. 136.1: Aspect-mood paradigms for bilı ‘buy’

a few exceptions not dealt with here, thecontexts of use for each form are the sameregardless of the meaning of the base. Thus,non.rls/pfv (also called basic form; cf.Schachter & Otanes 1972: 66 f.) is used in hy-pothetical and complement clauses and incommands (cf. (6) and (7)), non.rls/ipfv(Schachter & Otanes’ contemplated aspect) isused for future events (cf. (8)), rls/pfv(Schachter & Otanes’ perfective aspect) andrls/ipfv (Schachter & Otanes’ imperfectiveaspect) for past and present events, respec-tively (see illustrative text in 6).

In addition, there is a recent perfective for-mation (with prefix ka- � rdp1) which in-volves no voice marking and does not allowfor subjects (ang-phrases). Hence, all partici-pant expressions are either genitive or loca-tive marked:

(11) Ka-la-larorecent.pfv-rdp1-play

ko1.sg.poss

pastill

lamangonly

saloc

bata’.child

‘I have just finished playing with thechild.’

As may be immediately observable there aresome asymmetries in the paradigms in Tab.136.1. Strictly speaking, the realis patientvoice forms do not have a marker for voiceand, similarly, there is also no marker for ac-tor voice in the non.rls/ipfv form of the um-paradigm. These asymmetries are probablyrelevant for unravelling the diachronic devel-opment of the voice marking system. Theirrelevance for the synchronic analysis of thesystem is still in need of further exploration(see Himmelmann 1987: 157�171; Blake1988: 79 f. for some discussion). In this regardit may be noted that although most actionexpressions in Tagalog are voice and hencealso aspect-mood marked, it is possible to usebases denoting actions without further affixa-tion. In such uses action bases may be seman-tically undergoer oriented (often with a resul-tative connotation). For example, in (11) an-

tay could be replaced with the patient voiceform inantay.

(12) Antaywait

ko1.sg.poss

angspec

sagotanswer

mo.2.sg.poss

‘I wait for/expect your answer.’

But unaffixed bases may also occur in im-peratives with actor orientation, e.g. hintayka ‘you wait’ (which could also be renderedwith mag-hintay ka), and for denoting a stateof affairs without orientation:

(13) Iyakcry

angspec

sagotanswer

niya3.sg.poss

saloc

akin.dat.1.sg

‘His answer to me was crying/to cry.’

In this last usage, unaffixed bases are similarto gerunds, to which we now turn.

4.3. GerundsFor bases denoting a state of affairs it is pos-sible to derive a form which is neither voicemarked nor aspectually and modally in-flected. This is done by prefixing pag- to thebase according to the following correspon-dence rules which hold between actor voiceand gerundial forms (cf. Schachter &Otanes 1972: 160):

actor voice gerund

-um- pag-mag- pag-rdp2mang- paN-rdp2

Tab. 136.2: Gerund formation

Gerunds are further derivable from stative ex-pressions (cf. 5.1, prefix ma- is substituted bypagka-) and joint action expressions (cf. 5.2,prefix maki- replaced by pakiki-). In general,gerunds may not be used in predicate posi-tion, since they are not oriented towards oneof the participants of the state of affairs de-noted. Only in a clause such as pag-lu-luto’ng pagkain ang trabaho niya ‘ger-rdp2-cookgen food spec work 3.sg.poss (his job is




cooking food)’ may a gerund be used predi-catively (pagkain � which may also mean‘eating’ � is a semantically specialised ger-und from kain ‘eat’). Gerunds are most com-monly used in noun phrases, e.g. nag-umpisang pag-si-sigaw ‘rls.av-begin gen ger-rdp2-shout ((the turtle) began shouting/to shout)’and nang-galing sa pag-su-sugal niya ‘rls.av-come.from loc ger-rdp2-gamble 3.sg.poss((this) is due to his gambling)’, or in subordi-nate clauses (usually with a temporal mean-ing) pag-balik ni Gabby sa Pilipinas ... ‘ger-return gen.pn Gabby loc Philippines (whenGabby returned to the Philippines ...)’. Asshown by the preceding examples, all partici-pants involved in the state of affairs denotedby a gerund have to be expressed in genitiveor locative phrases.

A special perfective form of the gerund in-dicates that the event took place before thatof the main clause. It involves the prefix ka-(which may be optionally reduplicated) fol-lowing the general gerund formatives men-tioned above (cf. Schachter & Otanes1972: 161). Compare pag-punta ‘going’ withpag-(ka)-ka-punta ‘having gone’. Gerundsmay become the basis for further derivationsinvolving voice, aspect, and mood (see 5.2).

Formations with paN- without reduplica-tion regularly denote instruments used in thestate of affairs denoted by the base, e.g. pa-mutol ‘cutting instrument’ (< putol ‘cut’),panghampas ‘a whip’ (< hampas ‘whip’),pangkape ‘means for buying coffee’, etc. (cf.Bloomfield 1917: 224 f.).

5. Actor involvement

Although much less widely discussed, Taga-log morphology concerning the way an actoris involved in a given state of affairs is evenmore elaborate than the voice morphology.The basic split here is that between dynamicand stative or potentive formations (for sta-tives see also Drossard 1984: 64�72). The dy-namic forms are morphologically unmarked

av(‘buy’) av(‘sell’) pv lv cv

non.rls/pfv maka-bilı maka-pagbilı ma-bilı ma-bilh-an ma-i-bilınon.rls/ipfv maka-bıbilı maka-pagbıbilı ma-bıbilı ma-bıbilh-an ma-i-bıbilırls/pfv naka-bilı naka-pagbilı na-bilı na-bilh-an na-i-bilırls/ipfv naka-bıbilı naka-pagbıbilı na-bıbilı na-bıbilh-an na-i-bıbilı

Tab. 136.3: Potentive aspect/mood paradigms for bilı ‘purchase, sale’

and have been dealt with in 4. That is, anaction expression marked for voice (and as-pect and mood) generally implies a volitionalactor who is in full control of the action (themajor exception being some expressions forprocesses marked with -um- briefly men-tioned in 4.1).

5.1. Stative and potentiveStates of affairs which do not involve a con-trolling actor are expressed by a related butdifferent set of formations. Two different sce-narios have to be distinguished here. On theone hand, the state of affairs may be suchthat it excludes the involvement of an actorfor principled conceptual reasons. This is typ-ically the case for stative expressions such as‘be hungry’, ‘be angry’, ‘be adrift’, and thelike. On the other hand, the state of affairsmay be such that in principle it allows forcontrolling actors but in the specific instanceat hand the conceptually possible controllingactor is not in full control of the event. Thisis the case when someone happens to dosomething without having the intention to doit. Formations expressing this second pos-sibility are called potentive.

A completely regular correspondence ex-ists between dynamic and potentive forma-tions. That is, for each dynamic form there isa corresponding potentive form. The majorpotentive formatives are maka- for actorvoice and ma- for the undergoer voices. Fordetails compare Tab. 136.1 with Tab. 136.3.

The typical use of potentive forms is forinvoluntary actions:

(14) Biglasudden

niya-ng3.sg.poss-conn

na-bigkasrls.pot.pv-enunciation

iyon:dist

‘(Terrified) she suddenly exclaimedthis: …’

This includes actions done accidentally, i.e.the actor may be in control of the action butdid not really intend its outcome:




stat stat.lv stat.cv stat.av

non.rls/pfv ma-galit ka-galit-an i-ka-galit maka-galitnon.rls/ipfv ma-gagalit ka-gagalit-an i-ka-gagalit maka-gagalitrls/pfv na-galit kina-galit-an i-kina-galit naka-galitrls/ipfv na-gagalit kina-gagalit-an i-kina-gagalit naka-gagalit

Tab. 136.4: Voice and aspect/mood paradigms for stative bases (base galit ‘anger’)

(15) Na-i-lutorls.pot-cv-cooked

ko1.sg.poss

na.now

‘I happen to have cooked it already(by mistake).’

It also includes perceptions over which theactor (� experiencer) has no control as in:

(16) doondist.loc

aypm

na-kitarls.pot.pv-see

nila3.pl.poss

angspec

isa-ngone-conn

ma-lakı-ngstat-size-conn

higantegiant

‘there they saw a great giant ...’

In a second, somewhat different use poten-tive forms express the ability of an actor toperform the action in question:

(17) kungkungif

inyonginyo-ng2.pl.dat-conn

mapagtiisanma-pag-tiıs-anpot-ger-suffer-lv

iyaniyanthat

‘if you are able to endure this ...’

(18) atatand

hindıhindı’neg

makabarılmaka-barılpot.av-gun

sasaloc

kanya.kanya3.sg.dat

‘(The man got bitten by the ants) andwasn’t able to shoot at him.’

Turning now to stative expressions, these alsocome in four different voices, two of whichare formally identical to the potentive voiceforms. Because of the formal similarities, thesame labels have been chosen here for gloss-ing these stative voices. However, their syn-tax and semantics differ quite clearly fromthe potentive formations so that the labelsare not really indicative of their functions.

In the basic voice form for statives, whichis simply called stative here, the subject is atheme, i.e. an entity which is in, or currentlyis undergoing, a given state. The forms areidentical to the patient voice forms of the po-tentive paradigm, i.e. base plus prefix ma-.The forms of the stative actor voice are mor-phologically identical to the potentive actor

voice forms, i.e. base plus maka-. The stativelocative and conveyance voices are markedby the prefix ka- to which the basic voice af-fixes -an and i-, respectively, are added. SeeTab. 136.4 for an overview of the forms.

The only really productive formation is thebasic stative formation. Almost every Taga-log content word base can be prefixed withma- and then expresses a state:

(19) na-galitrls.stat-anger

siya3.sg

‘she was/got angry’

With bases such as basag ‘crack, break’,which allow both a state and an action read-ing, the form nabasag is ambiguous: It canmean ‘be in a broken state’ (stative) or ‘hap-pen to break/able to break’ (potentive patientvoice). In context, these readings are gen-erally distinguished by the presence of anovert actor expression in the potentive use(nabasag niya ‘s/he happened to break it/wasable to break it’).

The stative locative voice is common withbases expressing emotions. The subject ex-presses the person or thing at which the emo-tion is directed:

(20) kinagalitanin-ka-galit-anrls(ugr)-stat-anger-lv

siyasiya3.sg

ngnggen

Nanaynanaymother‘mother was angry with him/her’

Frequently, stative locative voice derivationstake on some more specialised meanings.Thus, kagalitan also means ‘to reprove, toscold, to rebuke’. In addition, stative locativevoice derivations are possible with a (rela-tively small) number of stative expressionswhich do not pertain to emotions. They thendenote the place at which a given state occurs(e.g. ka-matay-an ‘place where someonedied’, ka-hulug-an ‘place where someonefalls’).




The stative conveyance voice is also mostcommon with bases expressing emotions.Stative conveyance voice formations alwayshave the connotation of causation, that is,the subject specifies the reason for the emo-tion:

(21) ikinagalıti-in-ka-galitcv-rls(ugr)-stat-anger

niyaniya3.sg.poss

akoako1.sg

‘she got angry at me (I was the reasonfor her being angry)’

The stative conveyance voice is found witha somewhat broader range of bases than thestative locative voice. These include ikabasa’‘get wet on account of’, ikabagsak ‘fall on ac-count of’, ikatawa ‘laugh on account of’,ikaiyak ‘cry on account of’, etc.

The stative actor voice is very similar inmeaning to the stative conveyance voice sinceit also specifies the cause for a given state.But the two formations differ in their gram-mar and productivity. In the stative convey-ance voice construction (as in the stativelocative voice construction), the theme argu-ment (i.e. the one who experiences an emo-tion in the case of emotions) is grammaticallycoded as a genitive argument. In the stativeactor voice construction it is a locative argu-ment:

(22) lahatlahatall

ngnggen

kanyangkanya-ng3.sg.dat-conn

sabihinsabi-instatement-pv

ayaypm

nakagagalitnaka-rdp1-galitrls.stat.av-rdp1-anger

sasaloc

akinakin1.sg.dat‘everything he says irritates me’

The subject expression in the stative actorvoice construction usually refers to an inani-mate cause (some state of affairs or a thing).With regard to productivity, the stative actorvoice forms are the least common of all sta-tive formations and whenever they occur theyoften take on somewhat specialised meanings(thus makagalit is ‘irritate, antagonise, giveoffence’ rather than a plain ‘make angry’).Furthermore, the stative actor voice deriva-tions are often conventionalised in one of thefour aspect/mood forms, for example, naka-aawa’ ‘arousing pity, pitiable’ (< awa’ ‘mercy,compassion’), nakaka-lito (or naka-lılito)

‘confusing’ (< lito ‘confused, at a loss’), ornakaka-ganda (or naka-gaganda) ‘beautify-ing’ (< ganda ‘beauty’).

5.2. Further modes of actor involvement

The following prefixes mark further modes ofactor involvement:

� pa- is a causative prefix and compatiblewith all voice affixes (see McFarland1984 for ample exemplification and dis-cussion). Actor voice is used when thecauser is the subject, patient voice whenthe causee is the subject, e.g. p<in>a-talım niya ito ‘<rls(ugr)>caus-sharp3.sg.poss prox (he made this sharp)’ andpa-patul-in mo si Huan ng kugon ‘caus-cut-pv 2.sg.poss PN Juan gen species.of.grass (have Juan cut the cugon-weeds)’.Patient voice in non-causative construc-tions becomes conveyance voice in caus-ative ones. Compare the following twoexamples:

(23) Nanakaw-inrdp1-nakaw-inrdp1-stealing-pv

babaint

natinnatin1.pi.poss

angangspec

bangkaybangkaycorpse

ninigen.pn

Andrea?AndreaAndrea

‘Will we steal Andrea’s corpse?’

(24) Ipananakawi-pa-rdp1-nakawcv-caus-rdp1-stealing

babaint

ulı’ulı’again

sasaloc

atinatin1.pi.dat

angangspec

bangkaybangkaycorpse

ninigen.pn

Andrea?AndreaAndrea

‘Is (he) asking us to steal Andrea’scorpse again?’

� paki- indicates that the actor joins an on-going action (sociative). It may also beused to make a polite request or to indi-cate that something is done as a favour.Next to actor voice maki-, all undergoervoices are possible, for example:

(25) i-p<in>aki-hulogcv-<rls(ugr)>sociat-fall

nigen.pn

PedroPedro

angspec

akingdat.1.sg.conn

sulatletter

‘Pedro mailed my letter (along withhis)’




� si- only co-occurs with mag- and indi-cates plurality of (individual) actors, e.g.nag-si-ilag sila sa bayan ‘rls.av-pl-flee3.pl loc town (they all fled from thetown)’.

Further differences in the conceptualisationof an action are indicated by the prefixes pag-and paN-. As illustrated in 4.3, these prefixesare used to derive gerunds. Such gerunds,which do not have an inherent orientation,are also compatible with voice affixes. Thus,contrasting sets such as the following occur:p-um-utol ‘cut’, mag-putol ‘cut several things’,mamutol ‘cut selectively or in quantity’. Sim-ilar contrasts involving undergoer voices arerare (an example is tapakan ‘step on’ vs. pag-tatapakan ‘step on repeatedly’). The use ofvoice marked paN-derivatives is not verycommon and generally indicates intensive,distributive or repeated action, e.g. bumilı‘buy’ vs. namilı ‘shop’, humampas ‘hit with awhip’ vs. manghampas ‘whip people, go whip-ping’.

The major contrast is the one between-um- and mag- (cf. Blake 1925: 248 f.; Lopez1937: 46�49; Pittman 1966; Schachter &Otanes 1972: 292 f.; Drossard 1984: 87�92;Himmelmann 1987: 185�188). Often mag-indicates the greater frequency or intensity ofan action, cf. bumasa ‘read’ vs. magbasa ‘toread a lot/study’. A similar formation, i.e. theprefixing of mag- plus rdp2, is possible inprinciple for any base to indicate intensive orrepeated action (cf. Schachter & Otanes(1972: 337 f.); Bloomfield (1917: 237�239)specifies stress shifts which may also occur).mag- and -um- may even co-occur to indicatea high degree of intensity, cf. mag-um-aral‘study diligently’ or mag-s-um-igaw ‘shout(long and very loud)’. In other instances,the contrast seems to pertain to transitivity:t<um>ayo’ kami ‘<av>stand.upright 1.pe(we stood up)’ vs. nag-tayo’ kami ng bahay‘rls.av-stand.upright 1.pe gen house (weerected a house)’. Note that in the corre-sponding undergoer voices no pag- appears:i-t<in>ayo’ niya ang bahay ‘cv-<rls(ugr)>stand.upright 3.sg.poss spec house (heerected the house)’. In this type of example,the base denotes some kind of position ormotion, and the um-form denotes an actorwho moves himself, the mag-form an actorwho moves something. A similar contrast ex-ists with respect to bases denoting qualities,

e.g. um-init ‘become/get hot’ vs. mag-init‘make hot, heat’. Much quoted, but unique isthe contrast between bumilı ‘buy’ and mag-bilı ‘sell’.

Apart from such contrasting sets there arealso several bases which only allow voicemarking for pag-derived stems. For example,from bawal ‘prohibited’ neither *bumawalnor *ibawal may be derived, but only magba-wal and ipagbawal. These bases do not dis-play a common semantic or phonologicalfeature. Other examples are bilin ‘order, in-struction’, kanulo ‘betrayal’, lingkod ‘ser-vant’, etc. (cf. McFarland 1976 Appendix II;Himmelmann 1987: 151). Another group ofbases � which again does not exhibit a com-mon denominator � allows actor orientationonly with mag-, while undergoer orientationis possible without prior derivation. For ex-ample, from luto’ ‘cook’ *lumuto’ may not bederived but only magluto’, while the un-dergoer voice form is simply iluto’ or lutu’in.Other bases belonging to this group are dasal‘prayer’, hugas ‘wash’, punas ‘wipe off’, libıng‘burial’, bayad ‘payment’, kahoy ‘wood’, hu-bad ‘naked’, etc. (cf. Himmelmann 1987:179 f.).

6. Illustrative text

The standard Tagalog orthography is usedwith the modifications noted in 1.2.

Angspec

ulolfoolish

naconn

unggo’monkey

atand

angspec

ma-runongstat-knowledge

naconn

pagong.turtle

Minsanonce

angspec

pagongturtle

habangwhile

na-lı-ligo’rls.av-rdp1-swim

saloc

ilog,river

aypm

naka-kitarls.av.pot:rdp1-see

siya3.sg

nggen

isa-ngone-conn

puno-ng-sagingtree-conn-banana

naconn

l<um>u-lutang<av>rdp1-float

atand

t<in>a-tangay<rls(ugr)>rdp1-carry.off

nggen

agos.flow

H<in>ila<rls(ugr)>pull

niya3.sg.poss

saloc

pasigan,river

datapwatbut

hindıneg

niya3.sg.poss

ma-dalapot.pv-carry

saloc

lupa’.earth

Dahilcause

ditoloc.prox

t<in>awag<rls(ugr)>call

niya3.sg.poss

angspec

ka-ibig-anass-like-lv




niya-ng3.sg.poss-conn

unggo’monkey

atand

i-ni-alaycv-rls(ugr)-offer

niya3.sg.poss

angspec

ka-putolass-cut

nggen

puno-ng-sagingtree-conn-banana

kungif

i-ta-tanımcv-rdp1-plant

niya3.sg.poss

angspec


ka-parte.ass-part

T<um>ango’<av>nod

angspec

unggo’monkey

atand

h<in>ate’<rls(ugr)>divide

nila3.pl.poss

saloc

gitna’middle

mula’begin

saloc

mag-kabila-ngav-other.side-conn

duloend

angspec

punotree

nggen

saging.banana

In-angkınrls(ugr)-appropriate

nggen

unggo’monkey

angspec

ka-putolass-cut

naconn

mayexist

mgapl

dahon,leaf

dahilcause

saloc

panukala’plan

niya3.sg.poss

naconn

iyondist

aypm

tu-tubo’rdp1-grow

naconn

ma-butistat-good

kaysathan

ka-putolass-cut

naconn

wala-ngneg.exist-conn

dahon.leaf

Nangwhen

maka-raanav.stat-way

angspec

ilangsome:conn

araw,day

angspec

punotree

nggen

unggo’monkey

aypm

namatay,rls.stat:dead

yamangwhereas

angspec

saloc

pagongturtle

aypm

t<um>ubo’<av>grow

hangganguntil

saloc

mag-bunga.av-fruit

Angspec

mgapl

sagingbanana

aypm

na-hinog,rls.stat-ripe

datapwatbut

hindıneg

ma-akyatpot.pv-climb

nggen

pagong.turtle

Dahilcause

ditoloc.prox

t<in>awag<rls(ugr)>call

niya3.sg.poss

angspec


ka-ibig-angass-like-lv.conn

unggo’monkey

atand

i-ni-alaycv-rls(ugr)-offer

niya3.sg.poss

angspec

ila-ngfew-conn

bungafruit

nggen

sagingbanana

kungif

a-akyat-ınrdp1-climb-pv

niya3.sg.poss

angspec

puno’.tree

Angspec

unggo’monkey

aypm

um-akyatav-climb

atand

k<um>ain<av>eat

nggen

makakaya.utmost

S<in>abi<rls(ugr)>say

nggen

pagong:turtle

“Hulug-androp-lv

mo2.sg.poss

ako.”1.sg

Datapuwatbut

i-s<in>agotcv-<rls(ugr)>answer

nggen

unggo’:monkey

“Balatskin

manthough

atand

ma-linamnamstat-delicious

aypm

hindıneg

kita1.sg.poss:2.sg

hu-hulug-an.”rdp1-drop-lv

Angspec

pagongturtle

aypm

na-galitrls.stat-angry

atand

nag-sabugrls.av-scatter

siya3.sg

nggen

tinıkspine

saloc

paligıdsurroundings

nggen

puno’.tree

Nangwhen

l<um>ukso<av>jump

angspec

unggo’monkey

na-tinıkrls.stat-spine

siya.3.sg

P<in>ag-bintang-an<rls(ugr)>ger-suspect-lv

niya3.sg

angspec

pagongturtle

atand


h<in>anap<rls(ugr)>look.for

upangso.that

pa-rusah-ancaus-suffer-lv

niya.3.sg.poss

Na-hulirls.pot.pv-catch

niya3.sg.poss

angspec

pagongturtle

saloc

kabila’other.side

nggen

isa-ngone-conn

tuod.stump

S<in>abi<rls(ugr)>say

niya3.sg.poss

saloc

pagong:turtle

“kita1.di

aypm

akingdat.1.sg.conn

pa-ru-rusah-an.caus-rdp1-suffer-lv

Mamiliav:choose

ka2.sg

saloc

dalawa.two

Dikdik-ıncrush-pv

kita1.sg.poss:2.sg

saloc

lusongmortar

oor

lunur-indrown-pv

kita1.sg.poss:2.sg

saloc

ilog?”river

Angspec

ma-runongstat-knowledge

naconn

pagongturtle

aypm

nag-umpisarls.av-begin

nggen

pag-si-sigawger-rdp2-shout

atand

h<in>ilıng<rls(ugr)>request

niya3.sg.poss

saloc

unggo’monkey

na,conn

kungif

ma-a-are’,stat-rdp1-possible

aypm

dikdik-ıncrush-pv

siya3.sg

saloc

lusong.mortar

Datapwatbut

i-s<in>agotcv-<rls(ugr)>answer

nggen

unggo’:monkey

“I-bı-bigaycv-rdp1-give

ko1.sg.poss

saloc

iyodat.2.sg

angspec

pa-rusacaus-suffer

naconn

hindıneg

mo2.sg.poss

gusto.”liking

Atand

i-ni-hagiscv-rls(ugr)-throw

niya3.sg.poss

saloc

ilogriver




angspec

pagong.turtle

Nangwhen

d<um>apo’<av>alight

angspec

pagongturtle

saloc

tubigwater

aypm

nag-si-sigawrls.av-rdp2-shout

siya3.sg

atand

s<in>abi<rls(ugr)>say

niya3.sg.poss

saloc

unggo’:monkey

“Salamat,thank

ka-ibig-an.ass-like-lv

Itoprox

angspec

akingdat.1.sg.conn

tirah-an!”dwell-loc

Bloomfield’s (1917) translation:“Once upon a time, when the turtle was

swimming in the river, he saw a banana-treeadrift and being carried along by the current.He dragged it to the beach, but was not ableto carry it up to the solid ground. Thereforehe called his friend, the monkey, and offeredhim a half of the banana-tree, if he wouldplant his part for him. The monkey agreed,and they divided the banana-tree at the mid-dle, half-way from either end. The monkeytook the half which had leaves, because hethought it would grow better than the halfwhich had none.

When a few days had passed, the monkey’stree died, while that of the turtle grew untilit bore fruit. The bananas grew ripe, but theturtle could not climb for them. Therefore hecalled his friend, the monkey, and offeredhim some of the fruits of the banana, if hewould climb the tree. The monkey climbedup and ate for all he was worth.

Said the turtle: ‘Throw me some.’But the monkey answered: ‘Though sweet

the skins, I’d throw you none.’The turtle got angry and scattered spines

round the foot of the tree. When the monkeyjumped down, he landed on the spines. Hesuspected the turtle and looked for him, inorder to punish him. He found the turtle be-hind a stump.

Said he to the turtle: ‘I am going to punishyou. Choose between the two: shall I brayyou in a mortar or drown you in the river?’

The clever turtle began to shout andbegged the monkey, if it were possible, tobray him in a mortar.

But the monkey answered: ‘I shall give youthe punishment you don’t want.’

And he threw the turtle into the river.When the turtle arrived in the water, he set

up a shout and said to the monkey: ‘Thankyou, friend! This is my home.’” (Bloomfield1917: 16)

7. Uncommon abbreviations

av actor voicecv conveyance voicelv locative voicepm predicate markerpn proper nounpv patient voicesociat sociativespec specifier

8. References

Adams, Karen L. & Manaster-Ramer, Alexis(1988), “Some Questions of Topic/Focus Choice inTagalog”. Oceanic Linguistics XXVII, 79�101

Bader, Thomas & Werlen, Iwan & Wymann,Adrian (1994), Towards a Typology of Modality.Bern: Institut für Sprachwissenschaft (Arbeits-papier 32)

Blake, Barry J. (1988), “Tagalog and the Manila-Mt Isa Axis”. In: La Trobe Working Papers inLinguistics 1. Melbourne: La Trobe University,77�90

Blake, Frank R. (1906), “Expression of Case by theVerb in Tagalog”. Journal of the American OrientalSociety 27, 183�189

Blake, Frank R. (1925), A Grammar of the TagalogLanguage. New Haven: American Oriental Society(American Oriental Series 1)

Bloomfield, Leonard (1917), Tagalog Texts withGrammatical Analysis. Urbana: The Univ. of Illi-nois (University of Illinois Studies in Languageand Literature 3.2�4)

Capell, Arthur (1964), “Verbal Systems in Philip-pine Languages”. Philippine Journal of Science 93,231�249

Carrier-Duncan, Jill (1984), “Some Problems withProsodic Accounts of Reduplication”. In: Aronoff,Mark & Oehrle, Richard T. (eds.), Language,Sound, Structure. Studies in Phonology Presented toMorris Halle by His Teacher and Students. Cam-bridge/MA, London: MIT Press, 260�286

Constantino, Ernesto (1971), “Tagalog and otherMajor Languages of the Philippines”. In: Sebeok,Thomas A. (ed.), Current Trends in Linguistics, Vol.VIII.1. The Hague, Paris: Mouton, 112�154

Crowhurst, Megan J. (1998), “Um Infixation andPrefixation in Toba Batak”. Language 74, 590�

604




Cubar, Ernesto H. (1984), “From Borrowed Nounto Verb: A Study of Functional Shift in PhilippineLanguages”. In: Paz, Consuelo J. & del Corro, Ani-cia (eds.), Studies on Philippine Minor Languages.Quezon City: University of the Philippines, 119�

133

DeWolf, Charles M. (1988), “Voice in Austro-nesian Languages of Philippine Type: Passive,Ergative, or Neither?”. In: Shibatani (ed.), 143�

193

Drossard, Werner (1984), Das Tagalog alsRepräsentant des aktivischen Sprachbaus. Tübin-gen: Gunter Narr

Foley, William A. (1976), Comparative Syntax inAustronesian. Ph.D. dissertation, University ofCalifornia, Berkeley [UMI 77�4453]

Foley, William A. & van Valin, Robert (1984),Functional Syntax and Universal Grammar. Cam-bridge etc.: Cambridge Univ. Press

Gil, David (1993), “Tagalog Semantics”, BerkeleyLinguistic Society 19, 390�403

Gonzales, Andrew B. (1971), “The Tagalog NasalLigature: A Reinterpretation”. Philippine Journalof Linguistics 2.1, 28�43

Guzman, Videa P. de (1978), Syntactic Derivationof Tagalog Verbs, Honululu: Univ. Press of Hawaii(Oceanic Linguistics Special Publication 16)

Himmelmann, Nikolaus P. (1987), Morphosyntaxund Morphologie: Die Ausrichtungsaffixe im Taga-log. München: Fink (Studien zur TheoretischenLinguistik 8)

Himmelmann, Nikolaus P. (1991), The PhilippineChallenge to Universal Grammar. Köln: Institut fürSprachwissenschaft (Arbeitspapier 15)

Himmelmann, Nikolaus P. (to appear), “LexicalCategories and Voice in Tagalog”. In: Austin, Pe-ter & Musgrave, Simon (eds.), Voice and Gramma-tical Functions in Austronesian. Stanford: CSLI

Humboldt, Wilhelm von (1838), Über die KawiSprache, Vol. II, Berlin: Dümmler

Kroeger, Paul R. (1993), Phrase Structure andGrammatical Relations in Tagalog. Stanford: CSLI

Lehmann, Christian (1984), Der Relativsatz.Tübingen: Gunter Narr

Lemarechal, Alain (1982), “Semantisme des partiesdu discours et semantisme des relations”. Bulle-tin de la Societe de Linguistique de Paris 77.1, 1�

39

Lemarechal, Alain (1989), Les parties du discours:Semantique et syntaxe. Paris: P. U. F.

Llamzon, Teodoro A. (1973), “The Four FocusTransformations of Tagalog”. In: Gonzales, An-drew B. (ed.), Parangal kay Cecilio Lopez. QuezonCity: Linguistic Society of the Philippines, 168�

183

Llamzon, Teodoro A. (1976), Modern Tagalog: AFunctional-Structural Description. The Hague,Paris: Mouton (Janua Linguarum, Series Practica122)

Lopez, Cecilio (1937), “Preliminary Study of theAffixes in Tagalog”. In: Constantino, Ernesto(1977, ed.), Selected Writings of Cecilio Lopez inPhilippine Linguistics. Diliman: Univ. of the Philip-pines Press, 28�104

Lopez, Cecilio (1970), “On the Boak Tagalog ofthe Island of Marinduque”. The Archive (Univer-sity of the Philippines) 1.2, 1�53

Marantz, Alec (1982), “Re Reduplication”. Lin-guistic Inquiry 13, 435�482

Marre, Aristide (1901), “Grammaire tagalog, com-pose sur un nouveau plan”. Bijdragen tot de Taal-,Land- en Volkenkunden 53, 547�92

Matsuda French, Koleen (1988), Insights into Taga-log: Reduplication, Infixation, and Stress from Non-linear Phonology. Dallas: The Summer Institute ofLinguistics and The University of Texas at Arling-ton (Publications in Linguistics 84)

Matthews, P[eter] H. (1981), Syntax. Cambridge:Cambridge University Press

McFarland, Curtis D. (1976), A Provisional Classi-fication of Tagalog Verbs. Tokio: Institute for theStudy of Languages & Cultures of Asia & Africa(Monograph Series 8)

McFarland, Curtis D. (21983), A Linguistic Atlas ofthe Philippines. Manila: Linguistic Society of thePhilippines

McFarland, Curtis D. (1984), “Tagalog CausativeVerbs”. Philippine Journal of Linguistics 15, 19�

47

Müller, Friedrich (1882), Grundriss der Sprachwis-senschaft, II, Band II. Abt., Die Sprachen derSchlichthaarigen Rassen. Wien: Alfred Hölder

Naylor, Paz B. (1980), “Linking, Relation-Mark-ing, and Tagalog Syntax”. In: Naylor, Paz B. (ed.),Austronesian Syntax. Papers from the Second East-ern Conference on Austronesian Languages [AnnArbor 1976]. Ann Arbor: Center for South andSoutheast Asian Studies (Michigan Papers onSouth and Southeast Asia 15), 33�49

Naylor, Paz B. (1995), “Subject, Topic, and Taga-log Syntax”. In: Benett, David & Bynon, Theo-




dora & Hewitt, George B. (eds.), Subject, Voice andErgativity. London: School of Oriental and AfricanStudies, 161�201

Pittman, Richard (1966), “Tagalog -um- and -mag-:An Interim Report”. In: Papers in Philippine Lin-guistics 1, Canberra: Pacific Linguistics (SeriesA.8), 9�20

Ramos, Teresita V. (1971), Tagalog Structures.Honululu: Univ. Press of Hawaii

Ramos, Teresita V. (1974), The Case System of Ta-galog Verbs. Canberra: Pacific Linguistics (SeriesB.27)

Ramos, Teresita V. & Bautista, Maria L. S. (1986),Handbook of Tagalog Verbs: Inflections, Modes, As-pects. Honolulu: Univ. of Hawaii Press

Reid, Lawrence A. (1981), “Philippine Linguistics:The State of the Art: 1970�1980”. In: Hart, DonnW. (ed.), Philippine Studies: Political Science, Eco-nomics, and Linguistics. Northern Illinois Univer-sity, Center for Southeast Asian Studies OccasionalPaper 8, 223�230

Rubino, Carl R. G. (1998), “The MorphologicalRealization and Production of a NonprototypicalMorpheme: The Tagalog Derivational Clitic”. Lin-guistics 36, 1147�1166

Schachter, Paul (1976), “The Subject in PhilippineLanguages, Topic, Actor, Actor-Topic or Noneof the Above”. In: Li, Charles N. (ed.), Subjectand Topic. New York etc.: Academic Press, 491�

518

Schachter, Paul (1995), The Subject in Tagalog:Still None of the Above. Los Angeles: UCLA/De-partment of Linguistics (UCLA Occasional Papersin Linguistics 15)

137. Diyari (Pama-Nyungan)

1. Introduction2. Morphological type3. Parts of speech4. Nominal morphology5. Pronoun morphology6. Verb morphology7. Predicate determiner morphology8. Particles9. Post-inflectional morphology

10. Illustrative text11. Uncommon abbreviations12. References

Schachter, Paul & Otanes, Fe T. (1972), TagalogReference Grammar. Berkeley etc.: Univ. of Cali-fornia Press

Scheerer, Otto (1924), “On the Essential DifferenceBetween the Verbs of the European and the Philip-pine Languages”. Philippine Journal of Education7, 1�10

Shibatani, Masayoshi (1988), “Voice in PhilippineLanguages”. In: Shibatani (ed.), 85�142

Shibatani, Masayoshi (1988, ed.), Passive andVoice. Amsterdam, Philadelphia: Benjamins (Typo-logical Studies in Language 16)

Shkarban, Lisa I. (1995), Grammaticeskij stroj ta-gal’skovo jazyka, Moskva: Nauka

Soberano, Rosa (1980), The dialects of MarinduqueTagalog. Canberra: Pacific Linguistics (Series B.69)

Totanes, Sebastian de (41865), Arte de la lengua ta-gala y manual tagalog. Binondo: Miguel Sanchez yCa [11745]

Wolfenden, Elmer (1961), A Re-statement of Taga-log Grammar. Manila: SIL and Institute of Na-tional Language

Wolff, John U. (1993), “Why Roots Add the Af-fixes with which they Occur”. In: Reesink, Ger(ed.), Topics in Descriptive Austronesian Linguistics.Leiden: Vakgroep Talen en Culturen van Zuidoost-Azie en Oceanie, 217�244

Wolff, John U. with Centeno, M. T. C. & Rau, D.V. (1991), Pilipino through Self-Instruction, Vol. 1�

4. Ithaca: Cornell Southeast Asia Program

Nikolaus P. Himmelmann, Bochum(Germany)

1. Introduction

The Diyari language was traditionally spokenin the far north of South Australia, to theeast of Lake Eyre along the lower reaches ofCooper Creek � a mostly dry watercoursethat runs from western Queensland into LakeEyre. This is one of the driest and hottest re-gions of Australia with an average annualrainfall of about 100 mm (or 4 inches) andsummer temperatures regularly reaching 45



handbücher zur sprach- und kommunikations- wissenschaft filehandbücher zur sprach- und...

Documents