Rules vs. Lists

Rob Freeman lists at
Thu Jul 3 06:54:51 UTC 2008

On Thu, Jul 3, 2008 at 10:48 AM, David Tuggy <david_tuggy at> wrote:
> **Yes. Such "tumbling to" moments are actually, I am sure, quite common
> during the time when kids are learning 10 new words a day or however many it
> is that they absorb. I certainly can attest to them when learning a
> second/third/etc. language. But (a) They were not part of my language until
> I had them, and (b) once I'd had them they are on the way to being
> entrenched as conventional. When I encounter the same, or similar data again
> I will recognize it.

I think these "tumbling to" (ah-ha?) moments happen, on some level,
every time we say something new.

Indeed I think they are a model for how we say new things (to answer
your question "Why?")

Once something new has been said, it is on its way to being
conventionalized. Eventually the original "tumbling to" meaning may
become ossified and even replaced. I agree this conventionalization
aspect has been well modeled by CG. It is also important, but is
already being done well.

I won't question what CG tells us about the social, conventionalized
character of language. I'm only suggesting people consider this
"tumbling to" aspect to language. If it occurs, how many such new
generalizations might be made given a certain corpus of language
examples etc.

What it seeks to model are things which can be said. Whether something
which can be said, only becomes "part of my language" once I have said
it, is surely only a matter of definition.

Just to rewind and recap a little. The question at issue here is how
many generalizations/rules can be made about a list of examples. In
particular whether there can be more, many more than there are
examples. And the implications this might have for what can be said in
a language.


More information about the Funknet mailing list