lexical meanings in terms of linguistically expressed definitions, i.e. synonymous phrasal expressions. In addition,
suppose the principles by which sublexical units combine into word meanings are not the same as the principles by
which word meanings combine into phrase meanings. Then, even if the sublexical units com-posing a word meaning
could themselves be expressed as words, no phrase com-posed of those words could express what the original word
expresses. And of course, just as we have no conscious access to phonological primitives, we should not be able to
expound on word decomposition on the basis of raw intuition.
So here is the answer to Fodor's argument against definitions: although he has shown that lexical meanings cannot be
decomposed definitionally, he neglects the possibility that there are non-definitional forms of decomposition of the
sort found in phonology and physics. Since his alternative tonon-definitional decompositionis genetictransmissionof
the whole meanings ofquarkandfax, our choice ought to be clear.
An important question often raised about lexical decomposition is the justification of primitives. Where does
decomposition end? When do we know to stop? I don't think the parallel question in physics worries its practitioners
too much. The history of the last two hundred years is a continual quest to explain deeper and deeper regularities.
Semantics is just getting started; let's show some patience. In the meantime, we do the best we can to extract and
characterize relevant generalizations.
A quick example here might be helpful to illustrate both the non-definitional character of lexical decomposition and
the issue of primitives. A staple of lexical semantic analysis (going back at least to Gruber 1965, McCawley 1968, and
Lakoff1970) has beentheextractionofcausationas a significant componentofmanyverbmeanings. For instance, the
sentences in (1) are related to those in (2) by a common approximate paraphrase relation, shown in (3).
(1) a. The window broke.
b. The door opened.
c. The ball rolled down the hill.
d. Harry received aflower.
e. Harry died.
f. Harry decided to talk.
(2) a. Beth broke the window.
[≈Beth caused the window to break]
b. Beth opened the door.
[≈Beth caused the door to open]
c. Beth rolled the ball down the hill.
[≈Beth caused the ball to roll down the hill]