Croatian, along with other Slavic, and some Romance languages, has a t
ype of verb-fronting - the 'Long Head Movement' (LHM) construction - i
n which a non-finite verb raises to C0 across a finite auxiliary, in a
pparent contravention of the Head Movement Constraint. Rivero and Robe
rts have each argued on the basis of LHM-data that the Minimality cond
ition of the ECP should permit one head to cross another under certain
conditions. A detailed investigation of the properties of LHM in Croa
tian shows that the conclusions drawn by these authors with respect to
the ECP are inadequate in two respects. Firstly, the structural analy
sis for LHM sentences in Croatian assumed by Rivero and by Roberts is
incorrect. Clitics in Croatian are syntactically enclitic, right-adjoi
ning to C0. The auxiliary involved in LHM-constructions is a clitic fo
rm, hence is located in C0 at S-structure, and not lower down in the c
lause. Since the verb and clitic auxiliary in LHM occupy the same posi
tion, the Croatian data fail to support the view that the ECP must all
ow for one head to move across another. Secondly, although the notion
of 'last resort' is alluded to by these authors in connection with the
'triggering' of LHM in grammatical instances, the role of 'economy' i
n ruling out starred examples is not properly considered, rendering ar
gumentation with respect to ECP based on such (ungrammatical) examples
invalid. However, LHM does involve 'long movement' in the sense of 'a
cyclic incorporation'. We conclude that the correct definition of Mini
mality is one based on Baker's notion of 'distinctness' of heads.