This chapter complements the previous chapter on transitive motion verbs like 'push'. It deals with transitive motion verbs like 'pull' as in Lea pulled Gabriel out of the hole, where the theme (Gabriel) is moved away from a source (hole). In the English construction, the source is marked by a combination of two prepositions, out of. Other languages, however, feature the same preposition in motion-from contexts (1a) and in at-rest contexts (1b), e.g. Reunion Creole:
This kind of polysemous use of the preposition dan in Reunion Creole is strikingly different from the situation in its European base language French. In French, transitive motion-from constructions obligatorily show the ablative preposition de, as in tirer des pierres de la terre 'to pull stones out/from the ground', not the at-rest prepositions dans or sur (Les pierres sont dans la terre 'The stones are in the soil').
When we speak of at-rest situations, we refer – as in Chapter 82 – only to the spatial relations of containment ('in') and attachment ('at') and compare the motion-from constructions with these two kinds of at-rest constructions. The English sentence Gabriel is out of the hole may also be regarded as a kind of at-rest situation, but here it is implied that the location of Gabriel is a derived one, the result of precisely a motion out of some other location. So if we included all kinds of spatial relations, then even English would have the same marking for at-rest and motion-from.
As in Chapter 82, there are different means to express transitive motion-from situations, and languages may show several construction types.
|Special motion-from adposition||26||6||32|
|At-rest marking is used to express motion-from||12||10||22|
|Serial verb construction||4||3||7|
|Serial verb construction plus preposition||7||7||14|
In half of the APiCS languages, a special motion-from adposition (value 1) is used to mark the source from which the theme is removed.
The 22 languages with value 2 use their at-rest marking to express motion-from, as was already shown in ex. (1) from Reunion Creole. One other example comes from Fanakalo, where the (a) example shows the transitive motion-from construction and the (b) example shows an at-rest construction.
Here, the spatial relation 'near' occurs, which we consider very close to the spatial relation of attachment.
Interestingly, all French-based creoles in APiCS show this polysemous marking pattern, whereas there are only very few English-based languages with this value: Early Sranan, Krio, Ghanaian Pidgin English, Pichi, and marginally Tok Pisin.
Languages with the next two values express a transitive motion-from construction with serial verbs (value 3; 7 languages) or with serial verbs followed by a preposition (value 4; 14 languages). Within the languages with value 3, we find all four Portuguese-based creoles of the Gulf of Guinea.
The behaviour of transitive motion-from constructions strikingly parallels the behaviour of transitive motion-to constructions (Chapter 82), where the Gulf of Guinea creoles, too, constitute a compact linguistic area of serial verb marking.
Languages with value 4 (serial verb constructions with a preposition) are mostly found in the Guianas, West Africa, and the Pacific.
Here too, there is a clear parallel with transitive motion-to constructions (see Chapter 82). Sranan, Nengee, Tok Pisin, and Bislama mark both transitive motion-to and motion-from constructions with serial verb constructions plus a preposition.
When comparing the value distribution in the two chapters on transitive motion verbs (82 and 83), it is noteworthy that the values of serial verb constructions with and without a preposition (values 3 and 4), circumposition (value 5), and case marking (value 6) are very similiarly distributed.
The most striking differences in the distribution of the values concern the first two values "special motion-to/motion-from adposition" and "at-rest marking used to express motion-to/motion-from". Here we see very different proportions in both chapters. Compared to the 50 languages which treat 'push' verbs in the same way as the corresponding at-rest situation, only less than half of this number (22 languages) allow for this option with 'pull' verbs. Accordingly, there are three times as many languages with special motion-from adpositions than languages with special motion-to adpositions.
Unfortunately, there are no world-wide cross-linguistic data on transitive motion constructions. But considering the data in APiCS, there seems to be a correlation between intransitive and transitive motion marking (see Chapters 79, 80, and 81 for intransitive motion verbs). There is a general tendency for languages which mark intransitive motion-to and motion-from identically (lit. I go in the forest, I leave in the forest) also to mark transitive motion-to and motion-from identically (lit. I push s.o. in the hole, I pull s.o. in the hole). APiCS contains no chapter which compares transitive motion-to and motion-from constructions, analogous to Chapter 81 on intransitive motion constructions (see Chapter 81). But from the data we have, we can infer that languages which treat ‘pull’ constructions as at-rest constructions also mark ‘push’ constructions in the same way. Reunion Creole is an example of this. We already saw the transitive motion-from construction in ex. (1). The corresponding transitive motion-to construction (9a) is marked in the same way as are the intransitive counterparts (9b-c):
Already Boretzky (1983: 172ff.) identifies West African substrate languages with the same marking pattern. And there is some evidence that eastern Bantu languages also show this polysemy marking in transitive motion-to/-from constructions (see Michaelis 2008). See also Chapter 81 on intransitive motion-to/motion-from constructions, which likewise identifies substrate influence as the key to explaining the creole data.