Symmetrical voice, also known as Austronesian alignment or the Austronesian focus system, is a typologically unusual kind of morphosyntactic alignment in which "one argument can be marked as having a special relationship to the verb". This special relationship manifests itself as a voice affix on the verb that corresponds to the syntactic role of a noun within the clause, that is either marked for a particular grammatical case or is found in a privileged structural position within the clause or both.
There are two alignment types of languages with symmetrical voice: the Philippine type, which mostly retains the original system from Proto-Austronesian with four voices (or sometimes three), and the Indonesian type, which reduced them into only two voices.
The Philippine-type languages include languages of the Philippines, but is also found in Taiwan's Formosan languages, as well as in northern Borneo, northern Sulawesi, and Madagascar, and has been reconstructed for the ancestral Proto-Austronesian language. In the rest of the Malayo-Polynesian languages, including Proto-Oceanic, symmetrical voice was lost almost entirely.Beguš, Gašper. (2016). "The Origins of the Voice/Focus System in Austronesian". Presented at the 42nd Annual Meeting of the Berkeley Linguistics Society (BLS42).
The number of voices differs from language to language. While the majority sampled have four voices, it is possible to have as few as two voices, and as many as six voices. In the examples below, the voice affix on the verb appears in red text, while the subject, which the affix selects, appears in underlined bold italics.
Terminology The term Austronesian focus was widely used in early literature, but more scholars turn to the term voice recently because of the arguments against the term 'focus'.Himmelmann, N. P. (2002). Voice in western Austronesian: An update. In F. Wouk & M. Ross (Eds.), The History and Typology of western Austronesian voice systems (pp. 7-15). Canberra, ACT: Australian National University. On the other hand, Starosta argued that neither voice nor focus is correct and that it is a lexical derivation.Starosta, Stanley. (2002). Austronesian ‘Focus’ as Derivation: Evidence from Nominalization. Language and Linguistics, 3(2), 427-479.
Schachter (1987) proposed the word 'trigger', which has seen widespread use. As one source summarized, 'focus' and 'topic' do not mean what they mean in discourse (the essential piece of new information, and what is being talked about, respectively), but rather 'focus' is a kind of agreement, and the 'topic' is a noun phrase that agrees with the focus-marked verb. Thus using those terms for Austronesian/Philippine alignment is "misleading" and "it seems better to refer to this argument expression as the trigger, a term that reflects the fact that the semantic role of the argument in question triggers the choice of a verbal affix."Masumi Katagiri (2020) 'Tagalog'. In Tasaku Tsunoda (ed.) Mermaid Construction: A Compound-Predicate Construction with Biclausal Appearance. De Gruyter. P. 786.
Studies A number of studies focused on the typological perspective of Austronesian voice system.Hemmings, Charlotte. (2015). Kelabit Voice: Philippine‐Type, Indonesian‐Type or Something a Bit Different? Transactions of the Philological Society, 113(3), 383-405.Liao, Liao, H. C. (2011). Some morphosyntactic differences between Formosan and Philippine languages. Language and Linguistics, 12(4), 845-876. Some explored the semantic or pragmatic properties of Austronesian voice system.Kroeger, Paul. (2007). Morphosyntactic vs. morphosemantic functions of Indonesian –kan. In A. Zaenen et al. (Eds.), Architectures, Rules, and Preferences: Variations on Themes of Joan Bresnan (pp. 229-251).Huang, Shuan-fan. (2002). The pragmatics of focus in Tsou and Seediq. Language and Linguistics, 3(4), 665-694. Others contributed to the valence-changing morphology.Fortin, Catherine. (2003). Syntactic and Semantic Valence: Morphosyntactic Evidence from Minangkabau. In Proceedings of the Twenty-Ninth Annual Meeting of the Berkeley Linguistics Society (BLS 29).
Properties Agreement with the semantic role of the subject In languages that exhibit symmetrical voice, the voice affix on the main verb within the clause marks agreement with "the semantic role of the [subject]".Ross (2002, p. 20)
For example, the Actor Voice affix may agree only with agent nominal phrases. (The asterisk means that the sentence is ungrammatical for the intended meaning.)
Kapampangan
Tagalog
The sentences in (b) are ungrammatical because the patient nominal phrase is marked as the subject, even though the verb bears the Actor Voice infix. The sentences in (c) are ungrammatical because, instead of the agent nominal phrase, the location nominal phrase is marked as the subject.
The patient voice affix may agree only with patient nominal phrases.
Kapampangan
Tagalog
The sentences in (b) are ungrammatical because the agent nominal phrase is marked as the subject, even though the verb bears the patient voice affix. The sentences in (c) are ungrammatical because, instead of the patient nominal phrase, the location nominal phrase is marked as the subject.
The locative voice affix may agree only with location nominal phrases.
Kapampangan
Tagalog
The sentences in (b) are ungrammatical because the agent nominal phrase is marked as the subject, even though the verb bears the locative voice affix. The sentences in (c) are ungrammatical because, instead of the location nominal phrase, the patient nominal phrase is marked as the subject.
Types of semantic roles Across languages, the most common semantic roles with which the voice affixes may agree are agent, patient, location, instrument, and benefactee. In some languages, the voice affixes may also agree with semantic roles such as theme, goal, reason, and time. The set of semantic roles that may be borne by subjects in each language varies, and some affixes can agree with more than one semantic role.
Promotion direct to subject Languages that have symmetrical voice do not have a process that promotes an oblique argument to direct object. Oblique arguments are promoted directly to subject.
Tagalog
In the Tagalog examples above, the goal nominal phrase can either be an indirect object, as in (1), or a subject as in (2). However, it cannot become a direct object, or be marked with indirect case, as in (3). Verb forms, such as "nagpadalhan", which bear both an Actor Voice affix and a non-Actor Voice affix, do not exist in languages that have symmetrical voice.
The Tagalog examples contrast with the examplesTaken from Shiohara (2012)'s examples in (4a-b) on page 60, and in (12) on page 63. Glosses and translations modified for the Wikipedia article. from Indonesian below.
Indonesian
In the Indonesian examples, the goal nominal phrase can be the indirect object, as in (4), and the subject, as in (5). However, unlike in Tagalog, the goal nominal phrase in Indonesian can be a direct object, as in (6). The preposition kepada disappears in the presence of the applicative suffix -i, and the goal nominal phrase moves from sentence-final position to some verb-adjacent position. In addition, they can behave like regular direct objects and undergo processes such as passivisation, as in (5).
Proto-Austronesian examples The examples Taken from Blust (2013)'s Table 7.2 on page 439. Glosses and translations modified for the Wikipedia article. below are in Proto-Austronesian. Asterisks indicate a linguistic reconstruction. The voice affix on the verb appears in red text, while the subject, which the affix selects, appears in underlined bold italics. Four voices have been reconstructed for Proto-Austronesian: Actor Voice, Patient Voice, Locative Voice and Instrument Voice.
Proto-Austronesian
Formosan examples The data below come from Formosan, a geographic grouping of all Austronesian languages that belong outside of Malayo-Polynesian. The Formosan languages are primarily spoken in Taiwan.
Amis AmisTaken from Liu (2011)'s examples in (2.5) on page 27. Glosses and translation modified for the Wikipedia article. has four voices: Actor Voice, Patient Voice, Locative Voice, and Instrument Voice.
The direct case marker, which marks the subject in Amis, is ku.
Atayal While they both have the same number of voices, the two dialects of Atayal presented below do differ in the shape of the circumstantial voice prefix. In Mayrinax, the circumstantial voice prefix is si-, whereas in Squliq, it is s-.
Mayrinax MayrinaxTaken from Liu (2011)'s examples in (2.30) on page 44. Glosses and translation modified for the Wikipedia article. has four voices: Actor Voice, Patient Voice, Locative Voice, and Circumstantial Voice.
The circumstantial Voice prefix selects for benefactee and instrument subjects.
The direct case morpheme in Mayrinax is kuʔ.
Squliq SquliqTaken from Liu (2017)'s examples in (52) to (56). Glosses and translations modified for the Wikipedia article. has four voices: Actor Voice, Patient Voice, Locative Voice, and Circumstantial Voice.
The circumstantial voice prefix selects for benefactee and instrument subjects.
The direct case morpheme in Squliq is qu’.
Hla’alua Hla’aluaTaken from Pan (2012)'s examples in (3.16b), (3.23a), (3.32d) and (3.33a). Glosses and translation modified for the Wikipedia article.The orthography used in this subsection does not conform to the orthography used in Pan (2012) with respect to the consonant /ɬ/. Whereas Pan (2012) represents this sound as ‹lh›, this sound is represented here as ‹hl› (Pan (2012; page 50)). has three voices: Actor Voice, Patient Voice and Circumstantial Voice.
The circumstantial voice suffix selects for location and theme subjects.
While bound pronouns have a direct case form, nouns do not bear a special direct case marker for subjects in Hla’alua.
Kanakanavu KanakanavuTaken from Liu (2014)'s examples in (5a), (5c), (17a), and (20a). Glosses and translation modified for the Wikipedia article. has four voices: Actor Voice, Patient Voice, Locative Voice, and Instrument Voice.
The direct case morpheme, which optionally marks the subject in Kanakanavu, is sua.
Kavalan KavalanTaken from Lee (2016)'s examples in (24), and (25). Glosses and translations modified for the Wikipedia article. has three voices: Actor Voice, Patient Voice and Circumstantial Voice.
The circumstantial voice prefix selects for instrument and benefactee subjects.
The direct case morpheme, which marks the subject in Kavalan, is ya.
Paiwan PaiwanTaken from Ross and Teng (2005)'s examples in (2). Glosses and translations modified for the Wikipedia article. has four voices: Actor Voice, Patient Voice, Locative Voice, and Instrument Voice.
The direct case morpheme, which marks the subject in Paiwan, is a.
Pazeh Pazeh,Taken from Li (2000)'s examples in (22), (39), and (58), and Li (2002)'s example in (15). Glosses and translations modified for the Wikipedia article. which became extinct in 2010, had four voices: Actor Voice, Patient Voice, Locative Voice, and Instrument Voice.
The direct case morpheme, which marks the subject in Pazeh, is ki.
Puyuma PuyumaTaken from Aldridge (2015)'s examples in (7), and Cauquelin (1991)'s example on page 44. Glosses and translations modified for the Wikipedia article. has four voices: Actor Voice, Patient Voice, Locative Voice, and Circumstantial Voice.
The circumstantial voice suffix selects for benefactee and instrument subjects.
The direct case morpheme, which marks the subject in Puyuma, is na or i.
Seediq The two dialects of Seediq presented below each have a different number of voices. The direct case morpheme, which marks the subject in both dialects, is ka.
Tgdaya