The starting point of the research was the construction illustrated in (1), reported for West Flemish in Haegeman and Van Koppen (2012) and so far not described in the literature. In this pattern, what looks like a possessor constituent (Valère) is separated from the possessee (zijn auto) by an intervening adjunct (toen juist ).
(1) Het moest lukken dat Valère toen juist zijn auto in panne stond.
It so happened that Valere then just his car out of order was
The research undertaken has led to a number of separate lines of enquiry, as the data revealed the methodological complications that arise when looking at ‘non-canonical’ data whose status with respect to the grammar is unclear and the question arises how best to determine the acceptability of such data. This has led the researchers to look into a number of other non-canonical patterns, all relevant for the syntactic properties at issue (subjecthood, (sub)extraction, locality, (double) agreement, case checking etc.).
Six main lines of enquiry have developed which reflect this:
(i) the empirical basis of the research: non-canonical data and the question of defining the edge of acceptability;
(ii) the acceptability and the syntactic properties of Flemish External Possessors;
(iii) the acceptability of indirect passives in Dutch;
(iv) the acceptability and the syntactic properties of the catgap pattern, a pattern in English pattern where an unexpected gap appears as subject of a clause subordinate to a whose NP-relative clause, coreferential with the antecedent of that RC, as in (2).
(2) The police have rescued a cat whose owners thought was dead.
(v) the status and syntactic pattern of what looks like wh-raising illustrated in (3), with specific reference to the double agreement and the role of hyperactivity in syntax, a property also displayed in (4)
(3) These are proposals which are expected will generate huge income.
(4) This is the candidate whom everyone expects will win.
(v0 Given the relevance of extraction from subjects relevant for (1) and for (2) attention has also been paid on conditions on sub extraction.
In short, the project aims to deal with the issue of non-canonical data and the problem of how to best determine the acceptability of such data by looking closely into four such patterns with non-straightforward acceptability as well as with specific syntactic issues relevant for these patterns.