Hedde Zeijlstra of the University of Göttingen will give a Ling Lunch talk tomorrow titled “Upward Agree is superior.” in Oak 338 at 12:30 Mar. 4, 2014.
In Zeijlstra (2012), I argued that the syntactic operation Agree may only apply between a probe that carries an uninterpretable feature and a goal that carries a matching interpretable feature, where the goal is the closest potential goal that c-commands the probe. Similar conclusions have been reached by Wurmbrandt (2011, a.o.). This version of Agree has been dubbed Upward Agree (UA), as opposed to Downward Agree (DA), where the probe c-commands the goal. The proposal for UA has raised several criticisms, most notably by Preminger (2013). In short, Preminger stated that whereas certain instances of Agree should indeed be implemented in terms of UA (Negative Concord, Sequence of Tense), other instances of Agree should still be implemented in terms of DA. In this paper, I tentatively argue that the potential counter arguments rather show that Zeijlstra’s original UA proposal was not strong enough and that under an even more restrictive version of UA in combination with several particular assumptions about he syntax of case valuation the raised counterarguments may vanish. Moreover, under this revised analysis, a well-known instance of macroparametric variation would be fully explained. The novel empirical and theoretical contribution of the proposal is that DA can only take place if it is accompanied by instance of UA. I conclude the talk by showing that a particular instance of macro-parametric variation, observed by Baker (2008), follows from this proposal.