A passive construction is a special construction with a transitive verb (i.e. a verb that normally takes two arguments, an A and a P argument, as explained in Chapter 58) in which the argument that is normally the P is treated like the S in an intransitive clause, while the argument that is normally the A is either omitted or is treated like some oblique argument. An English example of an active (i.e. an ordinary, non-special construction) and a corresponding passive construction is given in (1).
In the following, the P argument will be called “object”, and the A and S arguments will be called “subject” for convenience.
In the WALS chapter on the passive construction (Siewierska 2005), there is only a simple binary presence/absence distinction. Somewhat less than half of the world’s languages have a passive construction, but in Europe, all languages have one. In the APiCS languages, presence of a passive is more common than absence, and to some extent this seems to be due to the influence of the European lexifiers. But the pidgins and creoles are very far from being simple copies of their lexifiers.
We distinguish four different types:
excl | shrd | all | ||
---|---|---|---|---|
Typical passive construction | 18 | 19 | 37 | |
Passive without verbal coding | 12 | 17 | 29 | |
Other atypical passive construction | 4 | 2 | 6 | |
Absence of passive construction | 23 | 0 | 23 | |
Representation: | 76 |
About half of the APiCS languages have a typical passive construction (value 1) with the following five features that characterize passives cross-linguistically (Siewierska 2005):
(i) it contrasts with another construction, the active;
(ii) the subject of the active corresponds to a non-obligatory oblique phrase of the passive or is not overtly expressed;
(iii) the object of the active corresponds to the subject of the passive;
(iv) the construction is pragmatically restricted in some way relative to the active;
(v) the construction displays some special verbal (morphological or auxiliary verb) coding.
In the English-based languages, the passive marker is usually the auxiliary get; passives with a ‘be’ verb are rare in our languages; an example is Spanish-based Papiamentu, where the passive can be formed with ser.
In some of the Portuguese-based languages, the passive is formed with the passive-participle suffix -du:
In the French-based languages, innovative verbs tend to be used as passive markers, such as trouve (< French trouver ‘find’) in Louisiana Creole and ganny (< French gagner ‘win, get’) in Seychelles Creole.
In the Southeast Asian languages, the passive is often restricted to adversity situations, where the patient is negatively affected by the action. This is also found with the overtly coded passive in Papiá Kristang, which uses the passive auxiliary toka (‘touch’):
Such a restriction of a new passive auxiliary is also reported for Mauritian Creole (Baker & Kriegel 2013), where gany (< French gagner ‘win, get’) is only used with a few verbs with meanings like ‘beat’ or ‘punish’.
In one APiCS language, the Tugu variety of Batavia Creole, a passive marker was borrowed from a substrate o adstrate language (di- from Malay):
For several of the languages, it is noted by the contributors that the passive construction is acrolectal, or more typically used in the written language, and often it is noted that it is not frequently used, which probably means that it is less common than in English.
A surprisingly high number of creole languages show a construction that is like the typical passive construction in all respects except (v), i.e. it has no verbal coding (auxiliary or affixal) (value 2). Such constructions are not common in the world’s languages (Haspelmath 1990 even claimed that they do not exist), but they occur widely in APiCS, in English-, French- and Portuguese-based languages, in the Atlantic and Indian Ocean (and often side by side with a typical passive construction):
This construction is unusual, and authors have sometimes been reluctant to call it a passive construction, though syntactically it behaves just like other passives (e.g. LaCharité & Wellington 1999 for Jamaican). It is similar to the phenomenon of ambitransitive verbs in English (e.g. break, which can mean ‘cause to break’ or ‘break by itself’), but intransitive counterparts cannot normally be formed from agent-oriented verbs like ‘plant’ or ‘sweep up’ (Levin & Hovav 1995). Thus, these cases are uncoded passives, not ambitransitives. The origin of these uncoded passives is not clear.
The third value subsumes a number of special constructions that are difficult to generalize over. We classify a construction as an atypical passive if it lacks one or two of the properties of typical passive constructions. For example, Singlish has a construction such as John give his boss scold ‘John was scolded by his boss’, which is clearly calqued on a similar Hokkien Chinese construction, where the active object becomes the subject, and the active subject becomes the object of the passive auxiliary, rather than an oblique-marked phrase. In Michif, there is a special inverse construction (for which Algonquian languages are famous) that must be used when the object is more topical than the subject, e.g. when it is a pronoun and the subject is a full NP (as in ‘The police overtook him’).
Many pidgins and creoles lack passives. This is true especially of the languages of the Pacific region, which fits with the world-wide picture in Siewierska (2005), where the most striking passive-less region is the New Guinea area. Instead of passive constructions, languages often use constructions with a generic third person plural form:
This construction type is well-known from European languages, too. Another non-passive type that can be used in some languages is a construction where the subject is simply omitted but the object remains an object, as reported for two of the Chabacano varieties:
Since Chabacano has the object-marking preposition kon, it is particularly clear here that the object’s status is not affected. The cross-linguistic distribution of this construction type (called desubjective in Haspelmath 1990) has not been studied yet.