The Elusive Passive

My first languages were English and Cebuano, a Philippine dialect in the Austronesian family. Although I studied English formally at school, I never did the same with Cebuano, so most of my engagement with it has been on a purely spoken level. Taking this class, however, piqued my interest in studying Cebuano more formally. The more I read and thought about Cebuano grammar, however, the more I started to doubt the existence of the passive in Cebuano. This brings me to the points I want to make in this post:

1. There is a distinction between passive structure and passive function or meaning.
2. The structure of passives is not universal.
3. There may not be a passive structure in Cebuano, even though passive meanings can be expressed. Although some linguists claim that some instantiations of the na- morpheme, prefixed to verbs in Cebuano, result in passive constructions, I propose a re-interpretation of na- clauses as adjectival rather than verbal passive.

Part I. The Passive: Structure vs. Function

In English, a verb is in “active voice” when the subject is the agent (i.e. the doer of the action) and the object is the patient (i.e. the recipient of the action, the one that undergoes a change in state as a result of the action). In “passive voice,” this order is inverted: The subject is the patient, and the object is the agent.

(1) Active:     He reads the book. [Subject = Agent = He, Object = Patient = the book]

     Passive: The book is read by him. [Subject = Patient = The book, Object = Agent = him]

Historically, passives have been classified as either periphrastic or strict morphological (Keenan & Dryer, 1981). Languages with periphrastic passives attach an auxiliary verb to the primary verb of the active construction, whereas languages with strict morphological passives inflect the primary verb of the active construction directly. This is evident in the contrast between the English periphrastic passive, which uses the auxiliary to be, in (1) above and the Latin morphological passive in (2) below.

(2) Active (Latin):    librum legit (He reads the book.)

     Passive (Latin): liber legitur (The book is read.)

So far, I have discussed two structural characteristics of the passive: (1) the inversion of agent and patient, and (2) the classification as either periphrastic or strict morphological passive. These structural characteristics, however, neglect the functional characteristics of the passive. By function, I mean the set of goals that the passive achieves. In English, we often put such goals in terms of “defocusing/demoting/detracting attention from the agent,” “emphasizing/promoting the patient,” “avoiding assigning blame/responsibility to the agent,” “conveying that the agent is not as important as the topic or patient” and the like.

Although the structure vs. function distinction in passives may seem trivial, I think it is actually quite easy to miss when it comes to English passives, since structure and function are so tightly coupled in English passive constructions. That is, when we hear an English sentence-utterance with a passive syntactical structure, we immediately assume the speaker is attempting to achieve passive functional goals like detracting attention from the agent, avoiding assigning blame, conveying that the agent is not very important, etc. And conversely, when we ourselves set out to detract attention from the agent, avoid assigning blame, etc. in English, we choose to put our message in a passive rather than active syntactical structure.

In other languages, passive structure and passive function are not as tightly coupled as they are in English. Put another way, other languages besides English make a clearer distinction between passive structure and passive function. In French, for example, it is possible to use a passive structure like that in (3) below, but it is also possible to achieve passive functions (i.e. convey passive meanings and passive goals) using active structures like those in (4) and (5) below. Because the passive structure in (3) comes off as overtly formal and literary in French, native speakers often avoid it by means of strategies like the impersonal pronoun on in (4) or the reflexive se in (5).

(3) Ce    livre   est                  lu            souvent.  (passive voice)

      This book  is-3SG-PRES read-PP   often.

      ‘This book is often read.’

(4)  On    lit                           ce   livre   souvent. (active voice)

      One  reads-3SG-PRES  this book    often

      ‘This book is often read.’

(5)  Ce    livre   se                   lit                        souvent. (active voice)

      This book  REFLEXIVE  read-3SG-PRES     often.

      ‘This book is often read.’

Part II. Non-Universality of Passive Structures

Failure to make the distinction between passive structure and passive function has led to mis-classification of some linguistic phenomena outside English as passive. Although there may be a trend nowadays to move away from an anglocentric analysis of linguistics, the fact that most of the literature in linguistics has examined English, combined with the assumption that a universal grammar exists, has misled some linguists into claiming all linguistic phenomena that exhibit structural similarities with the English passive are also functionally passive (Toyota, 2003).

Navajo, for example, predominantly uses SOV syntax marked by the prefix yi-, but also allows OSV syntax marked by the prefix bi-.  In contrast to SOV syntax, OSV syntax exhibits the sort of structural inversion of agent and patient that is found in English passives. However, the Navajo worldview precludes some instances of OSV syntax from mapping to a passive meaning.

Under the Navajo worldview, agents are on a hierarchy. Humans are perceived as higher on the hierarchy than animals, so humans can control animals but not vice versa. Therefore, one can say in Navajo, (6a) “The man kicked the horse”, but not (6b) “*The horse kicked the man” (active) nor “*The man was kicked by the horse” (passive).  Rather, the occasion of a horse kicking a man would have to be conveyed through an active construction that means, (7) “The man let the horse kick him (the man)” because humans can have causal efficacy over the motion of animals, but not vice versa.

(6) Navajo uses yi- for constructions involving controlling agents and non-controlling patients

(6a) hastiin łíí’       yiztał

        man    horse  it-it-kicked

      ‘The man kicked the horse.’

(6b) *łíí’     hastiin biztał

       horse  man    it-it-kicked

       *‘The horse kicked the man’ / *‘The man was kicked by the horse.’

(7) Navajo uses bi- for constructions involving controlling patients

     hastiin łíí’         biztał

     man    horse    it-it-kicked

     ‘The man let himself be kicked by the horse.’

As these Navajo sentences suggest, the conditions for passive in English may not generalize to all other languages. While the inversion of the usual agent-patient order may be sufficient for a sentence to count as passive in English, it is not sufficient for sentences in some other languages.

Part III. Cebuano Passives

As in Navajo, the deviation from the usual agent-first, patient-second order is not sufficient for an utterance in Cebuano to count as passive.

For some background, Cebuano is a dialect spoken on the Philippine island of Cebu. Sometimes, it is grouped with similar dialects from the middle third of the Philippines and collectively called Bisaya, Binisayan, or Visayan. Like other languages in the Austronesian family, Cebuano is verb-initial, with predominantly VSO word order, or in semantic terms, VAP order (Verb-Agent-Patient).

Two morphemes, gi- and na-, have been considered by linguists as candidates for passivization in Cebuano (Payne, 1994; Tanangkingsing & Huang, 2007). Here are some examples to give you a feel for these morphemes, with question marks next to translations that are still up for debate:

(8a) Gi-tangtang   namo  ang papel.

       Removed      we      the paper

       ‘We removed the paper’ OR ‘We are removing the paper’

(8b) Gi-tangtang   ang papel. 

       Removed      the   paper

       ‘(Someone) removed the paper’

(9a) Na-tangtang   namo   ang papel.

       Removed       we       the  paper

       ? ‘The paper was removed by us’ OR ‘We removed the paper’ OR ‘We managed to remove the paper.’

(9b) Na-tangtang   ang   papel.

       Removed     the   paper

       ? ‘The paper was removed’ (passive) OR ‘The paper came off’ (active)

My native intuition is in line with Tanangkingsing and Huang’s (2007) conclusion (from a study of spoken data) that gi- clauses are active. In Cebuano, gi- strongly implies effortful or intentional action from some agent, even if that agent is not explicitly named. The difference between the gi- clause in (8b) and the ni- clause in (9b) is that (8b) presupposes the existence of an agent who does the removing, whereas (9b) does not. The closest analogy I can think of to English is the distinction between (10a) ‘The tree was cut down’ and (10b) ‘The tree fell.’ In (10a), ‘cut down’ prompts you to imagine someone performing the action of cutting down something, but in (10b), ‘fell’ does not prompt you to imagine someone performing any action, since the tree could have fallen down for a number of reasons that do not involve the participation of a ‘someone,’ a human being.

I also agree with Tanangkingsing and Huang’s (2007) conclusion that na- clauses which name the agent, like in (9a), are active, since they make the agent accessible. However, I disagree with their claim that na- clauses that omit the agent, like in (9b), are passive. My intuition is that na-tangtang in (9b) is not “demoting the agent” or performing any of the passive functional goals I mentioned in Part I. Na-tangtang does not imply an oppositional hierarchy between agent and patient, where the agent has to be relegated to second place. Rather, na-tangtang is just describing the papel (‘paper’), as an adjective would. Therefore, I think it's possible that na-tangtang is an adjectival transformation of the verb root tangtang. Such a proposal would not be out of the ballpark for Cebuano, since the language allows verb-less sentences containing only adjectives and noun phrases. Of course, further analysis of spoken conversations and written text would need to be conducted to substantiate my hypothesis.

Conclusion and Discussion Points

The passive may not be a universal structure. For one thing, it is doubtful that all languages exhibit the passive. In fact, there is some reason to think the passive may have evolved in a subset of languages, including English, but not in others. Some linguists claim that the Proto-Indo-European languages originally only had an active and middle voice but no passive voice. The passive voice may have evolved from existing tense-aspect constructions when the language became more nominative-accusative (Toyota, 2003). Of the languages that are said to contain passive constructions, no single property is shared by all these constructions (Tanangkingsing & Huang, 2007). The lack of passives in some languages, and the lack of unifying necessary and sufficient conditions for passives among languages that are said to contain the passive, cast doubts on the role of the passive in a universal grammar.

Do you agree or do you think there is still a place for the passive in a universal grammar? Are there other instances in which an anglocentric linguistic analysis may not match your intuitions about linguistic phenomena (doesn’t have to be about passives) in another language you know? What are your thoughts on the passive? Can we communicate without it? If so, why is it still in our language?

*Note: Please feel free to make corrections to my non-English, non-Cebuano examples, which aren't in my first languages.

References

Keenan, E. L, & Dryer, M.S. (1981) “Passive in the world’s languages.”

Palmer, G. B. (1996). “Connecting Languages to World Views.” Toward a Theory of Cultural Linguistics. Austin, TX: University of Texas Press.
Payne, T.E. (1994). “The pragmatics of voice in a Philippine language: Actor-focus and goal-focus in Cebuano narrative.” In Givón, T. (Ed.), Voice and inversion (Volume 28, pp. 317-64). Philadelphia, PA: John Benjamins Publishing.

Tanangkingsing, M. & Shuanfan, H. (2007) “Cebuano Passives Revisited.” Oceanic Linguistics 46(2), 554-584.

Toyota J. (2003) “Anglocentric View and its Influence on Linguistics: a case of the passive voice.” Moenia 9, 51-73.

Voice (grammar). (n.d.). In Wikipedia. Retrieved November 18, 2014, from: http://en.wikipedia.org/wiki/Voice_(grammar))