#acl HlpLabGroup:read,write,delete,revert,admin All:read
#format wiki
#pragma section-numbers 2
#language en

= Construction grammar and psycholinguistics =
A discussion of construction grammar and its implications for pycholinguistics and vice versa.  Here are the readings:

== Short intro to construction grammar ==

* [http://www.princeton.edu/~adele/Publications_files/CaWCogLing-target%20article.pdf Overview of Adele Goldberg's Constructions at Work (30pp)]

== Some relevant psycholinguistic results ==
Read at least one of the following.  Even better would be to just familiarize yourself with the actual results of all 3, and skip or skim the background and discussions.

* [http://dx.doi.org/10.1515/COGL.2008.016 Zeschel, A. 2008. Lexical chunking effects in syntactic processing] NP objects in more frequent collocations are read faster, and the effect is stronger when finer-grained collocations are used.  Is this evidence that these fine-grained collocations are stored?  Does the experiment include the necessary controls to draw this conclusion?

* [http://www.sciencedirect.com/science?_ob=MImg&_imagekey=B6WCR-4T1SKGS-1-9&_cdi=6745&_user=483663&_pii=S0010028508000376&_orig=search&_coverDate=02%2F28%2F2009&_sk=999419998&view=c&wchp=dGLbVzz-zSkzS&md5=1329725e37cae1cff3be02e1f73b27ba&ie=/sdarticle.pdf Konopka, A. and K. Bock.  2009. Lexical or syntactic control of sentence formulation? Structural generalizations from idiom production] 
Idiomatic verb-particle constructions prime just as well as non-idiomatic ones.  Only syntactically "frozen" verb-particle phrases prime less.  What does this say about storage of fine-grained constructions vs coarser grained ones?

* [http://www.sciencedirect.com/science?_ob=MImg&_imagekey=B6WK4-4XKXRNJ-1-3&_cdi=6896&_user=483663&_pii=S0749596X09000965&_orig=browse&_coverDate=01%2F31%2F2010&_sk=999379998&view=c&wchp=dGLbVzb-zSkWb&md5=9c7167d5c0c6ad3868082fdabca5ade5&ie=/sdarticle.pdf Arnon, I. and N. Snider. 2010. More than words: Frequency effects for multi-word phrases] 
People respond to frequent 4-word phrases faster than less frequent ones, controlling for substring frequency.  Particularly relevant for debates in the CxG community about whether phrases of "sufficient frequency" are stored, a continuous frequency is a better fit than the best categorical predictor.  What kind of CxG would capture this?