This map shows the dominant order of lexical (nonpronominal) object and verb. As with Map 81A, the notion of object is defined semantically, as the P or most patient-like argument (see discussion under Map 81A) in a transitive clause. The primary types shown are languages which are OV (in which the object precedes the verb), illustrated by Turkish in (1a), and languages which are VO (in which the verb precedes the object), illustrated in (1b) by Gulf Arabic, the variety of colloquial Arabic spoken in Kuwait, Bahrain, Qatar, the United Arab Emirates, and eastern Saudi Arabia.
|Go to map|
|Object precedes verb (OV)||712|
|Object follows verb (VO)||705|
|Both orders with neither order dominant||101|
The third type is languages with both orders with neither order dominant; see “Determining Dominant Word Order”. A number of different subtypes of this type are discussed below. Note that the map does not distinguish languages in which only one order is possible and languages in which both orders are possible but one is dominant.
The map restricts attention to lexical noun phrases, ones consisting of a noun (plus possible modifiers), rather than objects consisting of just a pronoun. In some languages, pronominal objects occur in a different position from lexical objects. For example, in French, in which lexical objects normally follow the verb, pronominal objects normally precede the verb, as in (2).
‘I see him.’
Because lexical objects normally follow the verb in French, it is shown on the map as VO.
To a large extent, the SOV type shown on Map 81A corresponds to the OV type on this map. There are two other types on Map 81A that are OV, namely OVS and OSV, but these types are quite rare. Conversely, there are three types on Map 81A that correspond roughly to VO on this map, namely SVO, VSO, and VOS. There are a number of ways, however, in which these correspondences are not exact.
First, there are a number of languages which are shown as languages lacking a dominant order on Map 81A, but which are classifiable as OV or VO on this map. Some of these are languages in which one order of object and verb is dominant but which lack a dominant order of subject and verb in transitive clauses. The most common subtype of such languages consists of languages in which SVO is a common order in transitive clauses, but where VSO or VOS (or both) is also common. Syrian Arabic is an example of a language of this sort (Cowell 1964: 407, 411).
There are also languages in which OV is the dominant order and in which both SOV and OVS order are common so that they lack a dominant order on Map 81A. Macushi (Cariban; Brazil) is fairly rigidly OV, but SOV and OVS occur with about the same frequency (Abbott 1991: 25). In addition, there are languages in which the frequency of the two orders of object and verb depends on whether there is a lexical subject in the clause. For example, in Tonkawa (isolate; Texas), both SOV and SVO are common in clauses with both a lexical subject and a lexical object; but OV order is much more common in clauses lacking a lexical subject (based on my own text counts of texts in Hoijer 1972). Similarly, Yukulta (Tangkic; Queensland, Australia) is shown as SVO on Map 81A, but as OV on this map, since OV is reported to be preferred if there is no lexical subject while SVO order is preferred when there is a lexical subject (Keen 1983: 229).
There are also languages for which there is a dominant order both for the order of object and verb and for the order of subject and verb, but which do not have a dominant order for subject, object, and verb. Among these are languages where both VSO and VOS order are common but neither can be considered dominant. An example of such a language is Boumaa Fijian (Austronesian), illustrated in (3); only the context will determine which noun phrase is subject in a clause of the form Verb+NP+NP.
‘The old person saw the child.’ or
‘The child saw the old person.’
There are also many languages shown on this map that are not shown on Map 81A. These are languages for which it is clear from the available materials that the language is OV or that it is VO, but where the materials do not provide enough information to determine its type for Map 81A. Most languages shown on Map 81A with a specific word order (i.e. those not shown as lacking a dominant order) are shown on this map either as OV or as VO. An example of an exception is Paakantyi (Pama-Nyungan; New South Wales, Australia), which is SVO in clauses containing a lexical subject and a lexical object, but in which both OV and VO are common in clauses lacking a lexical subject (Hercus 1982: 236).
Languages in which neither OV nor VO is dominant fall into two sorts. On the one hand, there are languages with flexible word order where both orders are common and the choice is determined by extragrammatical factors. Many Australian languages, such as Ngandi (Gunwinyguan; Northern Territory, Australia; Heath 1978), are examples of this. A second class of language in which both OV and VO are common are languages in which word order is primarily determined syntactically, but in which there are competing OV and VO constructions. German is an instance of this, in that VO order is used in main clauses in which there is no auxiliary verb, as in (4a), while OV order is used in clauses with an auxiliary verb, as in (4b), and in subordinate clauses introduced by a subordinator, as in (4c).
‘Anna is drinking water.’
‘Anna has drunk water.’
‘Hans says that Anna is drinking water.’
A number of languages in Africa are similar to German in employing OV order in clauses containing auxiliaries, but VO order in clauses lacking an auxiliary. The example in (5) illustrates this for Kisi (Atlantic, Niger-Congo; Guinea): (5a), without an auxiliary verb, is SVO, while in (5b), with the present progressive auxiliary có, the verb follows the object.
‘The snake bit Saa.’
‘Fallah is sharpening the machete.’
Other instances in Africa, but far to the east of Kisi, include Nuer (Western Nilotic; Sudan; Crazzolara 1933), Dinka (Western Nilotic; Sudan; Nebel 1948) and Dongo (Ubangian, Niger-Congo; Democratic Republic of Congo; Tucker and Bryan 1966: 131).
Other instances of languages with syntactically determined order of object and verb are a number of Central Sudanic languages in eastern Africa, including the Moru-Ma'di languages, in which there are two constructions which can be broadly characterized as perfective and imperfective (or past and nonpast), in which the perfective construction is SVO, while the imperfective construction is SOV. The examples in (6) from Moru (Central Sudanic, Nilo-Saharan; Sudan) illustrate this.
‘I ate something.’
‘I am/was eating something.’
This contrast is not purely one of tense or aspect. For example in Avokaya, another Moru-Ma'di language, infinitival phrases are invariably OV while imperative clauses are invariably VO (Kilpatrick 1981: 98).
There are also languages in which the order of object and verb is partly sensitive to speech act type. For example, both Savi (Indic; Afghanistan; Buddruss 1967: 61-62) and Iraqw (Cushitic; Tanzania; Whiteley 1958: 64) are normally rigidly OV, but both allow VO in imperative clauses.
The distribution of OV order is similar to that described for SOV order in Chapter 81. OV predominates over much of Asia, except in the southeast. It also predominates in New Guinea, the exceptions being either languages along the north coast or on islands offshore; many of these exceptions are Austronesian. In Australia, OV predominates over VO but competes with languages in which neither OV nor VO order are dominant; even among those classified here as OV, the order of object and verb is generally relatively flexible. In the Americas, it is the dominant order outside two areas where VO predominates, Mesoamerica and the Pacific Northwest. In Africa, it is found to the west, north and northeast of the large area in which VO order is found, although the map is a bit misleading in that some of the areas in which OV order is found exhibit more genealogical diversity, so that in terms of genealogical groups, VO is less predominant in Africa than the map might suggest.
VO order is found in Europe and North Africa and among Semitic languages of the Middle East. It is the dominant type in Africa, though there are many OV languages around the periphery of the area in which VO is dominant. It is found in a large area stretching from China and Southeast Asia through Indonesia, the Philippines and the Pacific. Although it is the minority type in the Americas, there are two very well-defined areas that are almost exclusively VO, namely the Pacific Northwest (western Canada and the northwestern part of the continental United States) and Mesoamerica. Elsewhere in the Americas, VO order is found in a number of Algonquian languages of eastern Canada, in a number of languages of California, and sprinkled throughout South America, particularly among the languages further south.
Languages in which neither OV nor VO order is dominant are particularly common in Australia, and to a somewhat lesser extent, in North America. The Moru-Ma'di and Western Nilotic languages mentioned above, in which the choice between OV and VO is grammatically determined, form a clearly defined small area in eastern Africa.
The order of object and verb has received considerable attention because of the fact that a large number of other features are predictable from it, at least in a statistical sense (Greenberg 1963, Hawkins 1983, Dryer 1992). See Chapters 95, 96, and 97 for discussion. For example, OV languages tend to be postpositional (see Chapters 85 and 95), genitive before noun (see Chapter 86), adverb before verb, complementizer at end of clause, and standard-marker-adjective order in comparative clauses, while VO languages tend to exhibit the opposite orders. The patterns are sometimes more complex than this. For example, while VO languages almost exclusively place relative clauses after nouns, both orders of relative clause and noun are common among OV languages (see Chapter 96). In addition, there are some word order features which do not correlate with the order of object and verb. For example, contrary to some claims, the order of adjective and noun does not correlate with the order of object and verb (Dryer 1988a, 1992; see Chapter 97).
While it is often assumed in the literature that the order of object and verb has some privileged status among the various pairs of elements which correlate in order with each other, this assumption has not been supported. There is really no other good candidate among the various pairs of elements for such a privileged status. Perhaps the best alternative candidate would be adposition type (prepositions vs. postpositions); but many languages lack adpositions, yet still exhibit correlations among other pairs of elements. An alternative view is that no pair of elements has a privileged status; rather, there are just many pairs that correlate with each other, and the order of object and verb is just one of those pairs of elements.