MadSci Network: Chemistry 
Whats up with the Lewis dot structure of sulfate?
I Would like to have the structure of "covalent coordinate" molecules explained in terms of either molecular orbitals or hybridizations. Specifically, I am interested in the sulfate ion and carbond monoxide. I am not interested in hearing terms like "filling the valence shell" or "becoming like a noble gas". Thanks Josh, I remember you. You seem to be taking a strongly realistic view of things like orbitals and so forth, which is not entirely defensible (see, for example, Eric Scerri's "The Periodic Table and the Electron" in the NovDec 1997 American Scientist). An orbital is not something we would call an "observable" in quantum mechanics. Only the total electron density is observable, and if Lewis dot structures can help you predict the total electron density they're just as good as a molecular orbital picture. In fact, it is my opinion (for which I am in the process of constructing a formal argument) that Lewis structures now function as a shorthand for the valencebond picture of bonding. Second, if you want the full explanation in terms of "molecular orbitals or hybridizations," you really ought to read the references given in this MadSci answer. Incidentally, at least one of those references concluded that the Lewis dot picture (with octets and formal charges on sulfur) is a better fit to the electron density distribution than the "expanded valence" picture we get from hybridization theory (for which, of course, the best source is Pauling's The Nature of the Chemical Bond). But I'll see what I can do for you, in terms of both molecular orbital theory and valencebond theory  qualitatively, of course. A good general reference for this sort of thing is F.A. Cotton's classic text, Chemical Applications of Group Theory. If you want the full mathematical treatment of what is presented schematically below, you should work your way through a copy. Even then you won't have the mathematical equipment to do other than guess at the orbital energies...
Contents 
1. Carbon monoxide in terms of hybridization and resonance (i.e. valence bond) theory.CºO, being linear, is described by hybridization theory as having sphybridized atoms. Carbon begins with one electron in each of the following orbitals: sp_{1}, sp_{2}, p_{x}, p_{y}, shown schematically in Figure 1.Oxygen begins with six valence electrons, distributed (for example) thus: one in sp_{1}, two in sp_{2}, two in p_{x}, one in p_{y}, shown schematically in Figure 2.
However, the oneelectron orbitals on oxygen and carbon don't match up well, so we can fix things one of two ways:
Of course, valencebond theory requires that we mix these two models (which can in fact be stated mathematically); this is called "resonance" and is represented schematically in Figure 5. The true structure is neither of these models, but a combination, and includes contributions from electronegativity which shift electron density from carbon back to oxygen. The calculated charge density surface (HF/631G*) is shown in Figure 6.

2. Carbon monoxide in terms of molecular orbital theory.Qualitatively this is an extremely simple picture. The valence atomic orbitals on carbon and oxygen are, to a crude first approximation, identical. Since the molecular point group is C_{∞ v}, there are only two possible orbital symmetries: S and P. The s and p_{z} atomic orbitals have S symmetry, while the p_{x} and p_{y} have P symmetry.More strictly speaking, we need to realize that neither the C nor the O are at the molecule's center of symmetry, so that we must take symmetryadapted linear combinations of the atomic orbitals  which then turn out to be molecular orbitals right off. In point of fact, we don't usually work through the grouptheoretical analysis because simple visual combination gets us to the same place. And so, to work. There are two sets of orbitals of P symmetry, and they are mutually perpendicular so that they cannot overlap. They combine to form two pairs of degenerate molecular orbitals, one pair bonding (p) and the other antibonding (p*). These are shown in Figure 7. Figure 7 There are also two sets of orbitals of S symmetry, and they are not mutually perpendicular. We can combine them in two equivalent ways:
Figure 10 
3. The sulfate ion in terms of hybridization and resonance theory.
Figure 12
However, this is not the only possible valencebond picture! We can
entirely ignore pd pbonding (which is
not supported by quantum chemical calculations anyhow) and present a
picture which fully satisfies the Lewis octet rule. In this representation,
sulfur is still sp^{3} hybridized,
but dorbitals on sulfur do not participate in bonding. Instead,
bonding is semipolar (that is, a combination of covalent and ionic) and can
be represented by the Lewis structure of Figure 13. Of course, the +2
formal charge on sulfur is not quite a true representation. An
HF/631G* calculation calculated the charge on sulfur to be +1.77!
Induction, driven by the large separation of charge in Figure 13, pulls
electrons back toward sulfur so that the charge density surface for sulfate
ion (Figure 14) shows a moderation of the negative charges on the oxygen
atoms (the blue color is not positive charge, just "less
negative").

4. The sulfate ion in terms of molecular orbital theory.If one wants to be a purist, one can form the bonding framework of the sulfate ion exclusively from s and p orbitals on sulfur, interacting with a tetrahedral array of p orbitals on the oxygens. But it is just as valid to use Pauling's set of sp^{3} hybrid orbitals as a basis, as long as one combines them into sets which are compatible with the tetrahedral symmetry of the sulfate ion. These sets would have the same overall symmetry as the original atomic s, p_{x}, p_{y} and p_{z} orbitals, and combine with appropriate combinations of the p atomic orbitals on oxygen (that is, the p orbitals which point toward sulfur).Interestingly, sulfur d atomic orbitals (except possibly d_{z}2) do not appear to have the proper symmetry to form molecular orbitals in a tetrahedral molecule. The additional bonds generated by molecular orbital calculations are p bonds between the oxygen atoms, in which sulfur participates very little, if at all (for an example, see Figure 15). There are also several oxygen lonepair orbitals such as Figure 16.
Molecular orbital theory bears out the simple Lewis picture shown above in Figure 13; this is also the conclusion of A. Reed and P.v.R. Schleyer, Journal of the American Chemical Society 109, 736273 (1987), ibid. 112, 143445 (1990); E. Magnusson, ibid. 794051. 
5. The bottom lineDon't knock Lewis structures too much. They tell you almost as much, at considerably less expense in time, computing power and student frustration, as the more sophisticated and "correct" valencebond and molecularorbital models.The operative question, as always, is: does it correctly predict the observed electron density and chemical reactivity? For more on this, see "The Chemical Core of Chemistry" by Joachim Schummer, HYLE, 4, 129162 (1998). God never saw an orbital.

Try the links in the MadSci Library for more information on Chemistry.