International Computer Science  Institute

Construction Grammar Course


Taught by Professor Paul Kay and Professor Charles Fillmore

July 1-3, 2007

LSA.102P Introduction to Construction Grammar
1:30pm-3:30pm, Stanford University, Building 320 Room 105

July 1:
CJF presents a fairly informal survey of types of constructions, the general theme being: what kinds of phenomena does a "sign-based" construction grammar need to explain? Beginning with some simple-appearing observations on the nature of NPs, limited to Determination, Modification, and Complementation, we move on to a number of more complex structures - some of which don't at first seem complex at all. Is a NP like "the poor" really a case of an adjective being used as a noun? What are all the things we can do to calendric terms with the words "this", "next" and "last"? How are expressions indicating rate created and interpreted? Why can you negate comparative expressions with "no", as opposed to "not", and what is going on in the resulting phrases? What are the constraints on combining two numbers with "or" in making an estimate? What's to say about the fact that in "a mere five pages" we have a singular article but a plural noun? Why isn't there a definite article in a newspaper identification of "distinguished actor Charlton Heston"? And how do we account for the definiteness we sense in an unexpressed concept behind "We don't have the time"?

In too many cases we will not have a satisfying solution, but we will be able to ask important questions about what the possibilities might be.

Relevant Readings:
Fillmore's July 1 presentation (powerpoint file)

July 2:
CJF summarizes the results of the survey
PK introduces Sign Based Construction Grammar (SBSG), by first presenting an analysis of some argument structure constructions in a Cognitive Linguistics based constructional approach and then contrasting with this analysis an alternative formulation in SBSG.
Relevant Readings:
Chapter 3 (draft) of a forthcoming monograph on SBCG by (some ordering of) {C. Fillmore, P. Kay, L. Michaelis, I. Sag}, available here. Especially pp. 63-111. (Not to worry: page 63 is the first page.) Ideally, this should be read \bf before\ the first class.
This is the PPT presentation for the first hour on the second day. I had to leave the country before I had the chance to add the references to the literature on the "Correlative Conditional". I can provide that later on for anyone who asks. I'll be back "on duty" in August.

July 3:
PK continues introducing SBSG by contrasting the two styles of argument structure constructions, showing in particular how in SBSG lexical entries interact with \bf lexical class constructions\ to license full blown lexemes: e.g., the distinct lexemes \it bake\ in \it I baked a cake for you\ and \it bake\ in \it I baked you a cake\. (No additional reading.)


  • This paper, "Signs and Constructions", is the current version of the formal statement of the properties of sign-based construction grammar - chapter 3 of a forthcoming book. This is the main "heavy reading" for the course. Download or view in .pdf format
  • This paper, by Terry Langendoen, published in a Haj Ross festschrift, should be read in connection with CJF's first talk. It's a description of constraints on the pattern "Number or Number" in making quantity estimates in English. Download or view in .pdf format
  • For a pdf of the powerpoint Paul Kay presented with his lectures, click here
  • This 3-pager was intended as a contribution to the first chapter of The Book (not yet approved by the colleagues) and should be read in connection with CJF's presentations. The argument for such a "covert" construction is subtle and you ought to be able to examine all the claims. Download or view in .pdf format
  • Here's a powerpoint presentation of some work on the "calendar" expression constructions that'll get discussed in CJF's talk. It's in a long-abandoned notation (the "boxes-in-boxes" notation of early CxG), but the points should be recognizable. The second part of this presentation will be more directly relevant to our discussion.


Copyright © 2007 International Computer Science Institute. All Rights Reserved.