Notice that asymmetric and symmetric structures, after their generation by
unconstrained Merge, are utilized differently by CI: As we saw in Section 2,
the former yield various “endocentric,” d-structure semantics such as predicate-
argument structure and selection, whereas the latter yield “exocentric,” s-struc-
ture semantics such as topic-focus, operator-scope, theme-rheme, and so on.
Given our symmetry-driven syntax, we may make better sense of why this holds
at SEM: Merge-based SOs are “bare” and free from asymmetric representations
such as left-to-right ordering (X precedes Y) and projection (X projects over Y),
which were intrinsic to PSG-based/X-bar-theoretic syntax. SOs are no longer
universally endocentric, so we can reduce the notion of endocentricity to LIs’
unbalanced/asymmetric contributions to semantic interpretations at SEM, typical
of externally merged asymmetric SOs. In contrast, symmetrically organized SOs
show no such property, and hence their s-structure interpretations are not
dependent on any single LI (i.e., they are non-endocentric), so they exhibit
various discourse-related properties.
We can incorporate this observation into our generalization in (23) and sum
up the whole picture of symmetry-driven syntax as in (24).
(24) a. F-asymmetry b. F-symmetry
⇒introduced by EM ⇒ achieved by IM, head-
movement, and agreement
⇒formed before (b) ⇒formed after (a)
⇒exhibits endocentricity ⇒ exhibits no endocentricity
(exocentric)
⇒contributes to lexical, ⇒ contributes to discourse-related,
d-structure interpretation s-structure interpretation
(predicate-argument (quantifi cational, topic-focus, etc.)
structure, selection, etc.)
We specifi cally hypothesized that the DSC is the prime factor behind this overall
tendency in linguistic computation. Note again that this result can be achieved
only if we eliminate universal projection (2) and universal endocentricity (4),
which lends further support to the truly “bare” and “symmetric” formulation
of Merge in (5).
4 Equilibrium Intactness Condition
In the rest of this chapter, we will discuss some further consequences of
symmetry-driven syntax. In this section, we will propose another condition on
F-symmetry, the “Equilibrium Intactness Condition” (EIC), in an attempt to
capture the stable nature of F-symmetry.
As we saw above, the need for F-symmetry (the DSC) serves to trigger vari-
ous “symmetrizing” operations in syntactic derivation, such as IM, head-
movement, and feature agreement. Notice that the F-symmetric structures
created by these operations seem to be “stable” (as noted by Fukui 2011), and
20 Hiroki Narita and Naoki Fukui