We give here a new proof of the non-degeneracy of the funda
mental bilinear for
m for
me="mml1">method=retrieve&_udi=B6W9D-4K5HW64-1&_mathId=mml1&_user=10&_cdi=6680&_rdoc=5&_handle=V-WA-A-W-E-MsSAYZA-UUW-U-AAZCBUYAUV-AAZWYYYEUV-VBBDVADY-E-U&_acct=C000050221&_version=1&_userid=10&md5=ac1b5686e0b26a74ada5cbdae0fc39fe" title="Click to view the MathML source">Sn-
m-Quasi-Invariants
and for
m-Quasi-Invariants of classical Weyl groups. We also indicate how our approach can be extended to other Coxeter groups. This bilinear for
m plays a crucial role in the original proof [P. Etingof, V. Ginzburg, On
m-quasi-invariants of a Coxeter group, arXiv:
math.QA/0106175 v1, June 2001] that
m-Quasi-Invariants are a free
module over the invariants as well as in all subsequent proofs [Y. Berest, P. Etingof, V. Ginsburg, Cherednik algebras
and differential operators on quasi-invariants,
math.QA/0111005; A.
Garsia, N.
Wallach, So
me new applications of orbit har
monics, Sé
m. Lothar. Co
mbin. 50 (2005), Article B50j]. However, in previous literature this non-degeneracy was stated
and used without proof with reference to so
me deep results of Opda
m [E.M. Opda
m, So
me applications of shift operators, Invent. Math. 98 (1989) 1–18] on shift-differential operators. This result hinges on the validity of a deceptively si
mple identity on Dunkl operators which, at least in the
me="mml2">method=retrieve&_udi=B6W9D-4K5HW64-1&_mathId=mml2&_user=10&_cdi=6680&_rdoc=5&_handle=V-WA-A-W-E-MsSAYZA-UUW-U-AAZCBUYAUV-AAZWYYYEUV-VBBDVADY-E-U&_acct=C000050221&_version=1&_userid=10&md5=dac01759922a64dbce755d814c6a5903" title="Click to view the MathML source">Sn case, begs for an ele
mentary painless proof. An ele
mentary but by all
means not painless proof of this identity can be found in a paper of Dunkl
and Hanlon [C. Dunkl, P. Hanlon, Integrals of polyno
mials associated with tableaux
and the Garsia–Hai
man conjecture, Math. Z. 228 (1998) 537–567. 71]. Our proof here is not ele
mentary but hopefully it should be painless
and infor
mative.