But I really wanted to challenge your belief that "For first

language learners, irregularity in verb forms, for example,

is not a problem,...". As I've mentioned here before, this old
chestnut - popularised by Prof. Mario Pei among others - just
doesn't seem to be true. Let me quote a passage from "Lango"
I've posted here before:

".....the Turkish inflectional system is fairly intricate, but
  Turkish children normally master it well before the age
of two because it is completely regular and straightforward .
An oft-quoted illustration combines the noun "el" ("hand")
with the inflections "-im" (first person possessive), "-ler"
(plural) and "-de" (locative):

  el   "hand"

  elim   "my hand"

  elde   "in hand"

  elimde "in my hand"

  eller  "hands"

  ellerim   "my hands"

  ellerde   "in hands"

  ellerimde "in my hands"

  Similarly, relative clauses (i.e. those beginning "who,
  which or that" in English) are so straightforward in
Serbo-Croat that most Serbian and Croatian children
have likewise mastered them by the age of two.
Other languages also have grammatical features of
exemplary regularity which children learn to use without
difficulty. The international language committee will no
doubt look at all such instances in order to assemble
the best grammar from all sources.

  An interesting fact about the above two examples
is that their converse shows some of the worst
grammatical practice. Both Turkish relative clauses
and Serbo-Croat inflections are incoherent and
excessively complicated. The children of these nations
struggle to make sense of them, and do not normally
use them competently until about the age of five.

Why does he write this after first praising the Turkish system? Relative clauses in Turkish are about as easy as you can get.

Japanese relative clauses are even easier than that (just the infinitive plus the noun), by the way.