IMO, the situation can be summarized like this:

Actant A (somebody or something) influences Actant B's

I can see three semantic axes in this :
- how ? (by what means)
- is the influence oriented positively (the goal
being: action done) or negatively (goal: action not
- is the influence active (doing something in order to
make B act) or passive (not opposing)
- does the influence succeed or not ? or is it
undetermined ?

On the axe "how ?", I can see:
a. by physical means (forcing)
b. by moral means (persuading):
 b.1. referring to Actant A's moral authority
 b.2. referring to the authority of an institution
(society...) or a third person
 b.3. making Actant B feel that action is desirable
for himself.

In a, b.1 and b.2, Actant A tries to make Actant B act
against his will ; in b.3, Actant A changes Actant B's

So, causative, preventative, allowative or whatever
would simply be different combinations of the possible
values on these 4 axes. If we stick to the above
distinctions, that would give us: 5 x 2 x 2 x 3 = 60

(The 4th axis - result of the influence - looks very
much like aspect).

In case Actant A being something, rather than
somebody, we might have to reformulate some of the
definitions above. For ex, there would be no goal;
also, some possibilities would be absent

Ex: The snow nearly prevented him getting to the
station in time.
- How ? by physical means
- Influence oriented negatively
- Influence active
- Result not attained (nuance to be added: "nearly").

--- "Mark P. Line" <[log in to unmask]> wrote:
> Dan Sulani said:
> > I wonder why cultures don't seem to
> > feel a need (AFAIK) to grammatically mark
> prevention as well
> > as causation?
> In any reasonably parsimonious semantic description,
> every expression of
> prevention is also an expression of causation, while
> not every expression
> of causation is an expression of prevention --
> because it's easy to
> describe prevention in terms of causation and
> negation (both of which we'd
> be hard-put to do without in any semantic
> description). "Montezuma made
> Cortez eat possum." is not parsimoniously described
> as "Montezuma
> prevented Cortez from failing to eat possum.", and
> these two forms are
> even more pragmatically distinct than the oft-cited
> "kill" <--> "cause to
> die" pair.
> It's generally pretty easy to express prevention
> using whatever causative
> encodings you have in the language, so there's
> probably not often very
> much pressure for it to be grammaticalized to the
> same degree as simple
> causation.
> I also don't remember ever encountering a
> grammatical marker for PREVENT
> in a natlang or in any taxonomy of grammatical
> structures.
> -- Mark

Philippe Caquant

"High thoughts must have high language." (Aristophanes, Frogs)

Do you Yahoo!?
Friends.  Fun.  Try the all-new Yahoo! Messenger.