Chapter 12: Position of interrogative phrases in content questions
1. Interrogative phrases ⇫ ¶
In many languages, in particular in languages with verb-object order, the interrogative phrase (or “wh-phrase”) in content questions is normally or obligatorily fronted, i.e. it occurs in a special initial position in the clause that is different from the position that the corresponding non-interrogative expression would occupy. An example is English:
In other languages, they can occur in the position in which they would occur in the corresponding declaratives (“in situ”), or in another special position (e.g. a preverbal focus position). An example comes from Bislama, where the object interrogative phrase occurs in the ordinary object position:
2. The two types ⇫ ¶
Here we only distinguish two non-exclusive possibilities: initial (i.e. fronted) and non-initial (i.e. in situ, or in some other non-initial position), largely following Dryer (2005l). About half of the languages allow both possibilities.
| excl | shrd | all | ||
|---|---|---|---|---|
 | Interrogative phrase initial | 31 | 35 | 66 |
 | Interrogative phrase not initial | 10 | 35 | 45 |
| Representation: | 76 | |||
Two examples of initial position are given in (3a-b):
Three examples of non-initial position are given in (4a-c):
There may be further word order peculiarities in content questions, as in English, where the subject inverts with some verbs (as in 1b), but these hardly occur in our languages and are not taken into account here.
Note that “initial position” does not necessarily mean absolute initial position. Initial interrogative phrases may be preceded by highlighting particles or copulas in cleft constructions, as in example (5).
For this feature, we consider only interrogative phrases that do not normally occur initially in the clause, because with initially occurring elements (such as subjects), one cannot tell whether they are fronted. In example (6a), the interrogative phrase could be in situ or fronted, and it is only by analogy with (6b), where it is clearly fronted, that we may want to choose an analysis according to which (6a) is fronted as well.
It should be noted that in some languages that normally require fronting of interrogative phrases, occurrence of interrogative phrases in situ is possible in special echo questions (i.e. questions that copy part of an immediately preceding declarative sentence), e.g. in Afrikaans and English:
Because of their highly specialized usage, echo questions have been excluded for this feature, so Afrikaans is classified as allowing only fronted interrogative phrases.
3. Languages with both orders ⇫ ¶
Many languages allow both fronting and in situ occurrence of interrogative phrases. Such languages are Guadeloupean Creole, Lingala and Nigerian Pidgin:
There is often some semantic or pragmatic difference between the two construction types, but apart from echo questions (see above), such differences have been disregarded in our classification. For Bislama, Meyerhoff (2013) reports that in situ position (as in (2) above) is normal, while a question with fronted interrogative phrase is perceived as rude or aggressive. In Cape Verdean Creole of Santiago, in-situ position is especially used when the main verb is ‘be’:
In Creolese, in situ position often occurs with a suggested answer, or when the answer is considered to be obvious:
In some languages, different interrogative words behave differently. For instance, in Juba Arabic, wen ‘where?’ is always non-initial, whereas other interrogative pronouns may also be fronted (Manfredi & Petrollino 2013). Similarly, Santome bô ‘where’ obligatorily occurs in final position, whereas for other interrogative phrases, initial position is more common (Hagemeijer 2013). (This particular case can be explained etymologically, as bô was borrowed from Edo and is originally an interrogative copula.)
4. Geographical distribution of the types ⇫ ¶
In general, Atlantic pidgins and creoles show fronting of interrogative phrases, following the European model. In African languages, non-initial position is far more common (cf. Dryer 2005l), and the African pidgins and creoles also tend to show non-initial position (with the obvious exception of Afrikaans). In Asia, too, non-initial position is generally dominant in the indigenous languages, with the exception of Philippine languages, and this is reflected on the APiCS map, too. Thus, the position of interrogative phrases does seem to be susceptible to substrate influence in some areas, while in the Atlantic region the superstrate influence appears to be dominant. This is a puzzling pattern that requires further research.