precursor in the action domain of both humans and non-humans. If so, Merge
does have certain evolutionary continuity with a non-human, non-linguistic
function. In this weak sense, then, even Merge does not belong to FLN: FLN
is virtually an empty set.
In short, the FLN/FLB dichotomy is an illusion, arising from applying
different criteria to Merge and other components. In fact, Merge is uniquely
human only to the same extent that the CI and SM systems can be judged so,
and all of them are evolutionarily continuous with similar capacities of other
species. Merge is unique to human language, but then the CI and SM systems
are equally unique when these systems work together with Merge for specifi-
cally linguistic functions.
It seems reasonable to conclude that every component of language is equally
(non-)unique, and that the FLN/FLB dichotomy is a hindrance to our natu-
ral understanding of language and its evolution (see Boeckx 2013 fo r related
discussion).
6 Motor control origin of Merge
and Merge-only evolution
There have already been some discussions of the evolutionary continuity of Merge
with other biological traits, on how this computational capacity may have evolved
from preexisting non-linguistic functions shared by humans and non-humans.
Here I review the main proposals of Fujita (2014) with some amendment (see
also Boeckx and Fujita 2014 an d the references therein).
Both humans and non-humans have the capacity to combine physical objects
hierarchically and sequentially, as typically observed in tool using and tool mak-
ing. Chimpanzees’ nut cracking behavior is one classical example, but other
mammals, birds, and even insects are now known to be good tool users. What
deserves special mention is the metatool use by New Caledonian crows, where
the corvids ingeniously invent new ways of combining three distinct steps of
tool use to get food (Taylor et al. 2 010; more recently it has been reported
that they can combine as many as eight steps), though of course Morgan’s
Canon dictates us to ask whether these feats evidence the existence of the birds’
human-like insight (see Taylor et al. 201 2).
The hypothesis of the motor control origin of Merge maintains that Merge is
an exaptation of this object-combining function (call this Action Merge, adapt-
ing Greenfield’s (199 1, 1998) Action Gramm ar), which was later expanded to
manipulate abstract symbols in humans. The possible evolutionary links between
stone tool technology, language and cognition have often been discussed in
the field of cognitive archaeology, based on the observation that language and
tools share certain cognitive and neurological underpinnings (see Stout 2010
for e xample).
As noted, however, language is a complex trait, and each of its components
is very likely to have evolved from distinct precursors. The proposal here is that
the evolution of stone tool technology (as a form of Action Merge) should be
148 Koji Fujita