In Langlands' notes "On the classification of irreducible representations of real algebraic groups", available at the Langlands Digital Archive page here, Langlands gives a construction which is now referred to as "the local Langlands correspondence for real/complex groups".
What Langlands does in practice in this paper is the following. Let $K$ denote either the real numbers or a finite field extension of the real numbers (for example the complex numbers, or a field isomorphic to the complex numbers but with no preferred isomorphism). Let $G$ be a connected reductive group over $K$. Langlands defines two sets $\Pi(G)$ (infinitesimal equivalence classes of irreducible admissible representations of $G(K)$) and $\Phi(G)$ ("admissible" homomorphisms from the Weil group of $K$ into the $L$-group of $G$, modulo inner automorphisms). He then writes down a surjection from $\Pi(G)$ to $\Phi(G)$ with finite fibres, which he constructs in what is arguably a "completely canonical" way (I am not making a precise assertion here). Langlands proves that his surjection, or correspondence as it would now be called, satisfies a whole bunch of natural properties (see for example p44 of "Automorphic $L$-functions" by Borel, available here), although there are other properties that the correspondence has which are not listed there—for example I know a statement explaining the relationship between the Galois representation attached to a $\pi$ and the one attached to its contragredient, which Borel doesn't mention, but which Jeff Adams tells me is true, and there is another statement about how infinitesimal characters work which I've not seen in the literature either.
So this raises the following question: is it possible to write down a list of "natural properties" that one would expect the correspondence to have, and then, crucially, to check that Langlands' correspondence is the unique map with these properties? Uniqueness is the crucial thing—that's my question.
Note that the analogous question for $GL_n$ over a non-arch local field has been solved, the crucial buzz-word being "epsilon factors of pairs". It was hard work proving that at most one local Langlands correspondence had all the properties required of it—these properties are listed on page 2 of Harris-Taylor's book and it's a theorem of Henniart that they suffice. The Harris-Taylor theorem is that at least one map has the required properties, and the conclusion is that exactly one map does. My question is whether there is an analogue of Henniart's theorem for an arbitrary connected reductive group in the real/complex case.
Best Answer
Dear Kevin,
Here are some things that you know.
(1) Every non-tempered representation is a Langlands quotient of an induction of a non-tempered twist of a tempered rep'n on some Levi, and this description is canonical.
(2) Every tempered rep'n is a summand of the induction of a discrete series on some Levi.
(3) The discrete series for all groups were classified by Harish-Chandra.
Now Langlands's correspondence is (as you wrote) completely canonical: discrete series with fixed inf. char. lie in a single packet, and the parameter is determined from the inf. char. in a precise way.
All the summands of an induction of a discrete series rep'n are also declared to lie in a single packet. So all packet structure comes from steps (1) and (2).
The correspondence is compatible in a standard way with twisting, and with parabolic induction.
So:
If we give ourselves the axioms that discrete series correspond to irred. parameters, that the correspondence is compatible with twisting, that the correspondence is compatible with parabolic induction, and that the correspondence is compatible with formation of inf. chars., then putting it all together, it seems that we can determine step 1, then 2, then 3.
I don't know if this is what you would like, but it seems reasonable to me.
Why no need for epsilon-factor style complications: because there are no supercuspidals, so everything reduces to discrete series, which from the point of view of packets are described by their inf. chars. In the p-adic world this is just false: all the supercuspidals are disc. series, they have nothing analogous (at least in any simple way) to an inf. char., and one has to somehow identify them --- hence epsilon factors to the rescue.
[Added: A colleague pointed out to me that the claim above (and also discussed below in the exchange of comments with Victor Protsak) that the inf. char. serves to determine a discrete series L-packet is not true in general. It is true if the group $G$ is semi-simple, or if the fundamental Cartan subgroups (those which are compact mod their centre) are connected. But in general one also needs a compatible choice of central character to determine the $L$-packet. In Langlands's general description of a discrete series parameter, their are two pieces of data: $\mu$ and $\lambda_0$. The former is giving the inf. char., and the latter the central char.]