constraining equations question

[Stephen Wechsler]

Thanks, Ron, and a belated thanks to Ash and Helge.
Ron's answer is the one I was hoping to hear, for the particular
problem I'm working on.

best,
Steve

On Mon, May 2, 2011 at 11:31 AM, Ron Kaplan <ron.kaplan at post.harvard.edu> wrote:
> Steve,
>
>   I just came across your question.
>
>   The answer is that the constraining equation is not satisfied in the situation that you describe.  The f-structure corresponding to the root and the f-structure corresponding to the lower node are not the same f-structure, even though they share many properties. The constraining equations are evaluated on the f-structures in the minimal model for the collection of defining assertions, and in this case the minimal model for f does not include the ATT-VAL pair of the lower f-structure.
>
>  The shared properties are specified, presumably, by some attribute-by-attribute defining equations that leave out the particular ATT-VAL in your example. The typical result would be a subsumption--but not an equality--relation between the 2 structures.  But even that isn't necessarily the case: there can be other root-level defining equations that add attributes to the root f-structure that are not defined for the lower f-structure, and that would make it very obvious that the two structures are distinct.
>
>  I hope this helps to clarify the situation.
>
> Best regards--
>
> Ron
>
>
>