Well, I would tend to agree that a cartridge loading set too low is worse for the sound than one set too high. Some over-exuberence is preferable to constipation. But the question remains interesting, since from a technical point of view, it is well-understood that unloading an MC results in less-flat response, due to undamped resonances being allowed to flourish. That will indeed lead to the presence of "more information" - problem is, it will be spurious.
Maybe the difficulty lies in the relative lack of in-between loading choices offered by most phonostages. I have found it to be critical for fine-tuning to be able to experiment with multiple, fairly closely-spaced loading options in and around the "sweet-spot" range for a particular cartridge, so as to strike the correct balance between taut bass and sharp focus on the one hand, and easy dynamics and airy highs on the other. Sockets for removable resistor installation best facilitate this goal.
I also think it should be an instructive clue that 47k ohms represents a completely arbitrary value so far as MC design is concerned, having been established as a convention for MM applications long before MCs became popular. Any promotion of a 47k ohm standard for MC performance is in all likelihood as much a matter of default convenience dictated by history, as it is a considered choice based soley on sound. As I previously pointed out about the 47k value, not only is it arbitrary, it is well above the range where significant variations, or precision concerning exact values, have already ceased to make appreciable sonic differences. One may just as well advocate 25k ohms or 100k ohms - or 46k ohms - as 47k, since they will all be functionally similar when used with an MC that may have an internal impedance well below 50 ohms.
None of this, of course, is to say that in any particular set-up, 47k or 22k won't sound great, or won't sound better than the other options provided. This could be attributable a number of factors, such as some audiophile-approved speakers having a response that rolls off above 10KHz or so, overly-damped listening environments, or typical progressive hearing loss in older listeners, any of which could be roughly compensated for by MC unloading. There does exist a school of thought which holds that at least part of the reason why audiophiles adopted MCs in the first place, was due to the impressively airy response (overly so) and larger image size (but less physical) that resulted from running their new MCs into their old 47k MM phono preamps.
What this *is* to say however, is that IMHO, if you haven't tried many values beyond either 100 ohms or 47k (or 22k) ohms, then you haven't fully investigated your cartridge's performance potential. Most manufacturer presets or recommendations that offer 100 ohms as the common "loaded-down" setting choice are positioning the bar too low, resulting in sound which is suffocated. If you can, I would suggest getting a bunch of inexpensive 1/4 watt resistors (should cost just spare change each one) to try out in values of about 150, 200, 300, 500, 750, 1,000, 1,500, 2,000, 3,000, and 5,000 ohms as an educational range of options to get started in, and doing substitution comparisions, home in on the value range that provides the most natural-sounding performance. (A wide range of reference recordings should be used for the tests, but it is very helpful to include well-recorded examples of solo voice to aid in determining proper perspective of image focusing and size.)
What you are looking for is a value that represents the upper range for maintaining solidity, specificity, and definition, coinciding with the lower range for maintaining openess, liveliness, and bloom (you can use 100 ohms and 47k ohms as opposite extreme references if you already have them). Below the optimum range the sound will be stuffy and sterile; above, it will be amorphous and sloppy. The extreme high overtones should be neither closed-down, nor unrealistically exagerated (we're not talking about brightness here - the difference occcurs above that range); listen especially for coherent transient articulation to aid in determining this.
Once the right value range has been identified, get more resistors clustered around that value, in increments as small as 10-30 ohms for values below 500, to 100-200 ohms for values in the thousands, and repeat the auditioning process. Once the optimum value for your set-up, music, and preferences has been nailed down, replace the garden-variety resistors with premium ones of the same value to achieve best transparency and neutrality. (Naturally, all of the preceding assumes no step-up transformer is being used.)
I realize that what I am recommending here is at best a pain the kiester, at worst totally impractical or even impossible for some. However, I have come to accept that persistance in this area pays crucial dividends in the ultimate truthfulness and believability of phono sound, after having lived for years with 47k ohm loading as my only reference. It actually took undergoing this process to wean me away from the accustomed sound of running my MCs unloaded, and convince me I wasn't really missing anything that was actually supposed to be there - not to mention coming to appreciate what I was gaining. The fact that I could hear the differences and learn from them, given that my phono rig is hardly state of the art, points to even greater importance for optimization of loading in the best systems. To my mind, if you try this and find that unloaded still works best - and if you voiced your system around that unloaded MC sound - you can either go with what's most enjoyable and say to hell with accuracy, or it might be time to reconsider the wisdom of skewing the path toward synergy way back at the source, and maybe try repositioning the speakers or something to bring it all back together *with* the cartridge properly loaded. Just a thought, just my opinion... :-)
