Inclaimingthatchildrenlearngrammar soeasilybydoingitintuitivelyrather thanconsciously, critics alsomissanother
important point: in an information-theoretic sense, children and linguists are facing thevery sameproblem. I suspect
that this is the part of the argument that peoplefind most difficult. So let me try to make it more concrete, to show
how it works out in practice.
Suppose wefind thatin the (hypothetical)language Urtish, certain directobjects are marked dative case and others are
marked accusative. The problem for the child as well as for the linguist is tofind out which are which. Of course
children don't know anything explicit about case marking:“dative,”“accusative,”and even“case”are justourlabels.
Children only have to put the right affix in the right place, without knowing why, and perhaps even without knowing
thatthey have done so. But they do have to get it right.
Sowelinguistscanask howchildrenmanagetodothis. Onewaymightbethattheysimplymemorizetherightcasefor
each verb, and this may indeed be the right solution; in a theoretical grammar we would call this“lexically marked
quirky case.”But suppose when we introduce children to new verbs, they automatically use the right case:
(2) Experimenter:Look! This is blizzing. Tell me what I just did.
Child:You blizzed the cake-os. [where -osis the case-marking in question]Now we have to discover what clues the child used. Are they semantic, phonological, or pragmatic? If semantic, are
theybasedonthenatureoftheaction,theobjectactedon,theagent's intention,etc.,etc.?Thetheoreticalgrammar will
have to be formulated to capture this distinction. Or suppose children invariably use accusative case on thefirst try,
even when the correct case is dative. Then the linguist might say accusative is the“default”case, and look for the
systematicity or lack thereof behind deviation from the default. Havingfigured this out, then we can ask: Why did the
child look forthatclue rather than something else? Such questions are familiar to every practicing linguist. When we
find these questions arising in every corner of every language (as we do), the Paradox of Language Acquisition
becomes almost second nature, a fact of daily research life.
One of my favorite illustrations of the Paradox is the following set of sentences,first pointed out by Jeffrey Gruber
(1965).
(3) a. Every acorn grew into an oak.
b. Every oak grew out of an acorn.
c. An oak grew out of every acorn.
d. *An acorn grew into every oak.What is striking about this paradig mis its asy m metry. In (3a–c), we understand each acorn to correspond to an oak,
that is,everyquantifies overan. But (3d) has