I am *not* a Libertarian.
At least not in the sense argued against by Daniel C. Dennett
in chapter 4 his book Freedom Evolves.

I've known of Daniel Dennett ever since I read The Mind's I co-written with Douglas Hofstadter -- and made of commented texts by yet other authors. The titles of his books interested me greatly, and he has been in my reading list for a long time, although somehow, never on top of it until very recently. Then I saw the title of this latest book of his, and I immediately bought it from amazon. I haven't learnt much from the book, and I didn't expect to, either. I expected DCD to make a correct account of the notion of free will, and the way it is related to the dynamics of our existence at multiple levels of abstraction and feedback. Up to now (middle of chapter 4), I haven't been disappointed, though there are a few remarks I'd like to make.

In the context of the philosophical debate about free will

,
libertarians are people who believe in a metaphysical free will

as an add-on to deterministic material causation, in a way that is but
a rehash of immaterial souls in the guise
of modern philosophy and quantum physics.
I like the term moral levitation

used by DCD to describe their stance:
that free will

some kind of magic whim-machine
that floats above the universe and introduces random arbitrary choices
out of the blue.
The philosophical libertarians

are really animists who seek
a self-contradictory concept that be both materially observable yet immaterial.
They do not even understand the very notion of material world:
the world of all things directly or indirectly observable
through physical interaction.
If it's observable, then it's material.
Or maybe they don't understand the notion of a law of nature:
if it has any kind of behavioral pattern,
then this pattern is called a law.
If it has no kind of pattern, then it's random,
and there is nothing else to say about it.
Or maybe they just want something magic,
something to relieve them from having to think and understand.
Or maybe they want to identify some source of randomness or other
as the soul

of individuals --
whereas as Dennett will no doubt argue it in the rest of the book,
the soul is to be found precisely in the deterministic

patterns
that constitute the very character of an individual,
to the exception of this randomness.

Actually, what libertarians

completely miss out
is the notion of relevance.
The utterly unspeakable is utterly irrelevant.
But to be fair, it seems that DCD is unable to articulate it either,
though what he says amounts to that.
Determinism is irrelevant to the nature of the universe,
because it is a feature of models of our universe
that isn't intrinsic to the universe itself.
We can know the universe but through interaction.
Any modelling in terms of something lower-level than interaction
is but a matter of convenience for symbol-manipulation;
it doesn't reflect the structure of the universe,
but provides a tool for the mind to manipulate more easily;
if it reflects the structure of anything,
it is the structure of the minds that build the models.

For instance, mathematicians wondered for a long time whether
the the continuum hypothesis

was true.
Well, they found out that the hypothesis was irrelevant to most mathematics.
It's a feature of particular set theoretical models of mathematics,
that some may possess, that some may not possess,
and that may even be undefined in other models.
We can build all the usual structures of higher mathematics
without relying on the continuum hypothesis,
and we can even build all of them without resorting to set theory at all.
For instance, we can build mathematics on top of such a formalism
as category theory, where the CH doesn't even have a meaning.
My dad, a mathematician bent on the aesthetics of plane geometry,
taught me the meta-level concept of intrinsicness of a mathematical concept:
alignment is intrinsic to the projective geometry,
but not the choice of the hyperplane at infinity;
the origin is not intrinsic to an affine plane,
but to an arbitrary coordinate system chosen on it;
distance is intrinsic in euclidian geometry but not in affine geometry, etc.
What matters in a mathematical structure is its intrinsic properties.
Sometimes, it is easier to explore a structure
and to prove intrinsic properties of it
by introducing arbitrary intermediate objects
such as a coordinate system and a metrics;
but these intermediate objects are features of the demonstration
and are not intrinsic to the structure being explored.

Well, determinism is not intrinsic to the universe. For every non-deterministic model of the universe, we can introduce a deterministic model of it that is indistinguishable by any observation from internal interaction (by introducing an explicit source of pseudo-randomness to resolve indeterminacies); and conversely, for every deterministic model of the universe, we can introduce a non-deterministic model of it that is indistinguishable by any observation from internal interaction (by conflating any states that are undistinguishable through interaction). Moreover, instead of modelling the universe as a sequence of states or a graph of possible transitions that may be either deterministic or indeterministic, we may model it in very different ways; for instance, we may model it in terms of some modal logic relative to various human-observable and human-induced events that may take place -- and indeed, such representation is much more intrinsic to our relationship to the universe than any representation in terms of discrete particules, deterministic or not, though the paradigms are mutually expressible one in the other.

As a metaphor, in computer science, regular languages
may be represented by deterministic finite automata,
non-deterministic finite automata, or regular expressions;
and determinism is not intrinsic to languages themselves,
but to the choice of representation.
DFAs and NFAs are good for implementing recognizers
for strings in the language, or reasoning about the language,
each kind of representation having its advantages and disadvantages;
but regular expressions are the preferred way for humans to specify
such languages -- and are often extended toward specifying
more than just recognition of regular languages.
To claim that a language itself is deterministic

would be an utter misunderstanding of the issue,
a deep confusion between abstraction levels:
the concept of determinism

just doesn't apply
to languages, only to *some* models of it.

As for non-news, I am still a libertarian in the juridic sense: a proprietarist or a volontaryist, if you prefer. Thus I find it quite amusing that on p.6, DCD introduces an anti-libertarian proviso because he sees all too well how natural it is to deduce property from creation, yet has no argument against it: if he had, I'm sure he'd have included a footnote with a bibliographic reference. Oh well, nobody's perfect. I'm still very glad this book was written; at least I have a reasonable thing to cite when debating on the topic of free will.

__Post-Scriptum__:
Actually, DCD does understand the irrelevance of determinism.
For instance, at one place, he suggests that
it wouldn't make any visible difference should
the universe could be indeterministic in even days
and deterministic in odd days.
(Unhapily, I forgot to write down the page where he makes this suggestion,
and can't find it back.)
Other passages also display more concisely the same understanding:
p. 88 "assuming determinism is true, *or false,*
will not help him find the needle in this haystack."
p. 306 "(deterministic or indeterministic -- it makes no difference)"
But DCD fails to conceptualize the generic notion of irrelevance.
Or maybe he can conceptualize it, but specifically avoids invoking it,
preferring to argue the particular case.