Wednesday 12 November – 18:00 to 19:00 – Lecture Theatre 1, Appleton Tower
Robert Truswell (email@example.com) will be presenting on some joint work carried out with Nikolas Gisborne. It explores further some of the peculiar aspects of relative clauses that are currently an area of intense inquiry.
“This talk aims to make links between typological patterns and their theoretical explanations. I will focus in particular on the prospects for explaining one particularly unusual typological pattern, which we call “parallel evolution”.
Noun phrases like (1), where the relative clause has a wh-phrase in [Spec,CP], are found in roughly 1/2 of modern Indo-European languages, but only 1/40 of other modern languages.
(1) The person [[with whom] I saw John __ ]
In other words, wh-relatives are an Indo-European thing, more or less. Normally, if a phenomenon is specific to a language family in this way, we glimpse a diachronic explanation: properties of Proto-Indo-European (PIE) may be responsible for the skewed typological distribution of this construction. For instance, if PIE had constructions like (1), we would probably not be surprised if a lot of its daughter languages did too.
However, very few attested early IE languages (from more than a couple of thousand years ago) have wh-relatives like (1), so PIE almost certainly didn’t have them either. This means that wh-relatives have developed time and again, in parallel, across Indo-European, but not elsewhere. This is strange: typical models of syntactic change rely on random low-level reanalyses during acquisition, and the distribution of wh-relatives isn’t random.
Our aim is to relate this pattern to two factors: first, PIE must have some relevant distinguishing feature, or else the fact that wh-relatives are so concentrated in IE would be a mystery. We identify a likely candidate, a particular type of “correlative” construction. Second, certain types of reanalysis must be more likely than others, so that grammars are more likely to change in some directions than others. We identify certain biases operative in language acquisition that may plausibly lie behind this skew. These two ingredients, we hope, are all we need to explain the unusual typological distribution of wh-relatives.
Finally, this is work in progress, and the topic is proving to be a very rich source of research questions. We will do our best to show what it looks like at the relatively early stages of a project like this, and identify open questions and methodologies that could be used to address them.”
Entry free for members, £1 for non-members.
As usual we will be heading to Greenmantle afterwards, We hope you will join us.
RSVP via the Facebook Page.