Where did tracking error go?

Tracking error in pivoted arms as opposed to linear arms didn't go anywhere of course, it's just that, as with direct-drive vs. belt-drive, there are reasons why one method has generally prevailed in the home audio market, and those reasons have to do with factors other than what most resembles an inverse process for how LPs are cut, because there are advantages and difficulties to each method.

I don't agree that absence of skating forces is a bigger deal than incorrect azimuth. Variously adjusting anti-skate on my table produces relatively little change in sound quality. Even though the counterforce can't be 100% correct across the record's entire surface, I think it balances out at a low enough net force not to be a real problem for sound quality or record longevity. On the other hand, slightly changing stylus alignment angles can affect the sound more, especially with line-contact tip shapes.

It seems to me another potentially big advantage to linear-tracking would be the possibility of basically eliminating the tonearm (and any horizontal pivot point) altogether, but with the exception of the Souther most linear-tracking designs haven't pursued this goal for whatever reasons.

I didn't say anything against linear arms, or in preference of pivoted ones. If we're talking about the theoretical ideal for how a cartridge ought to traverse a record, of course the answer is linearly. I just don't feel that skating forces in particular are all that big a deal -- that they can be adequately compensated for, and don't affect the sound all that much. Your anecdote about tracking force seems intriguing, but I don't think it qualifies as definitive evidence on the effects of skating forces, or in choosing which type of arm to go with, due to uncontrolled variables. To me, the weakest points of any tonearm design are probably their resonances, followed by their bearings -- and then their geometry -- and as a breed, linear arms don't solve the first two problems any better than do pivoted ones, if even as well. I think there's always a trade-off involved either way you go, but that both approaches can be made to work acceptably well if not perfectly.

BTW, meant to correct this days ago and forgot, but the variable tracking angle error -- as compared with the constant cutting angle -- that's intrinsic to pivoted arms and absent from linear ones isn't the "azimuth", as was implied above (it's simply the tracking angle error). As I'm sure everyone really knows but may have temporarily spaced on, azimuth is the vertical perpendicularity of the stylus in relation to the plane of the record surface in line with the cantilever, and is adjusted by rotating the headshell or tonearm about its longitudinal axis, something that's not different for the two types of arms. Couldn't let that one stand for posterity you know...

That terminology? It's just an advertising man talking...