Explaining microvariation in P-dropping varieties

Explaining microvariation in P-dropping varieties

Certain syntactic constructions are standard (obligatory or optional) in some languages, and non-standard but possible in others. Especially where they are non-standard, it can go unnoticed that apparently similar constructions in varieties of the same language are in fact different, and cross-linguistic parallels can be overlooked because they are absent in the standard variety. Such is the case with preposition-dropping. The relevant phenomenon is illustrated in (1)–(5), where it is optional in the English varieties (but licensed under different conditions in the two), more or less obligatory in Greek (where including the overt preposition has marked pragmatic effects) and obligatory in Urdu and Haitian Creole French, where there is no overt counterpart:

(1) Does this train go Canterbury West? (Kent)

(2) Swim the end and back. (Liverpool, Biggs 2015: 53)

(3) Pao (ston) kinimatografo I.go (to.the.ACC) cinema.ACC

‘I am going to the movies.’ (Greek, Ioannidou & Den Dikken 2006: 1)

(4) Khala Karachi aaen gii

aunt.NOM Karachi.OBL.LOC come will

‘Aunt is going to come to Karachi.’ (Urdu, Sana Kidwai, p.c.)

(5) Timoun yo al Mache Pòspyewo

children DEF.PL go Market Post-Pierrot

‘The children have gone to the Post-Pierrot Market.’

(Haitian Creole French, DeGraff 2007: 122)

It is reported in varieties across Europe and beyond, particularly contact and non-standard varieties and appears to be a good contender for what Kortmann & Szmrecsanyi (2004) call ‘vernacular universals’. This paper focusses on the preposition-dropping of Southeast England, which differs in certain ways from that of Northwest England and has similarities with other European varieties. I show that a combination of existing mechanisms, when applied to constructions at a micro level rather than at a whole-language level, can successfully explain the properties of SE English P-dropping as set out in (a)–(e) below, as well as why Northwest English varieties are more permissive.

The construction in SE English is restricted to the following contexts. Only (a) is also true for (varieties of) NW English, while all are true for Greek:

a. The complement must be a Goal;

b. The location must be familiar or ‘anaphoric’ (in the sense that home refers to one’s own home);

c. The determiner must be omitted;

d. The verb is semantically weak (usually go);

e. The construction must be directional.

Myler (2013) and Biggs (2015) both argue that P-drop constructions have full PP and DP structure with a null case-marking element (TO or k). In both varieties, the determiner is usually overt, and so the question of whether there is full DP structure doesn’t arise. Conversely, Hall (2018) and Gehrke & Lekakou (2012) argue for pseudoincorporation of a bare nominal (i.e. no further structure in the DP or PP). It is far from clear, however, that the standard pseudoincorporation tests (e.g. Dayal 2011) hold in English, and there is no obvious reason for the restriction to locations rather than the more familiar direct object noun incorporation. Furthermore, there is some evidence for the existence of full DP structure: not only are the typical nouns in this construction place names, normally thought of as DPs, but occasionally, particularly in contrastive contexts, an overt determiner is permitted (Shall I go these services?). If there is full structure, this accounts for (a) above: the complement is a Goal, which can only be assigned by a lexical P to or its equivalent. We now turn to properties (b)–(e), characteristics of P-dropping comparable to Greek and romance varieties. (b) and (c) result from the treatment of the location as a Bare Singular Count Noun functioning as an ‘activity-naming predicate’ (Stvan 2009) or ‘institutionalised location’ (Gehrke & Lekakou 2012). The noun raises to D (e.g. Longobardi 1994) and they effectively function like place names. (Note also that in some varieties, the argument must be a place name.) (d) and (e) can be explained by the notion of ‘conflation’. Talmy (1985 et seq.) argued that in verb-framed languages like Spanish, Path (or directionality) is conflated with the verb and manner of motion must be expressed with an adjunct:

(6) La botella entró a la cueva flotando

the bottle moved.in to the cave floating

‘The bottle floated into the cave.’ (Talmy 1985: 69)

The English equivalent may include a manner of motion verb (float) because Path is not conflated and is expressed in a ‘satellite’ adjunct. If a manner of motion verb is used in a verb-framed language, it must express location rather than direction. Gehrke & Lekakou (2012) claim that prepositions can only be omitted in verb-framed languages. However, note that the same conditions of verb-framed languages described above also pertain to P-drop constructions in English. Compare *The bottle floated bridge, which is not grammatical, but to the extent that it is interpretable, it expresses location rather than direction. We therefore have a straightforward explanation of the restriction to directional rather than stative verbs, and prohibition on manner of motion verbs: when P is null, Path conflates with the verb to license the null P (as per Den Dikken 2010).