If a speaker can be bi-wired, it can also, by definition, be bi-amped. Since the Polk powered towers cannot be bi-amped, they therefore cannot be bi-wired either.
Polk's powered towers have two sets of binding posts which would seem to indicate that they can be bi-wired and bi-amped. However, this is not the case. The top set of binding posts feeds the midrange and tweeter while the lower set of binding posts feeds the speaker-level input of the built-in subwoofer amplifier.
1. speaker-level connection to either set of terminals with the binding post jumper in place
2. speaker-level connection to the top set of terminals with a line-level connection (RCA) from the "sub out" of the receiver/preamp to the sub in on the speakers
3. speaker-level connection to the top set of terminals with a line-level connection (RCA) from the left and right channel preouts of the receiver/preamp to the sub in on the speakers
The "Bi-Wire" Hook-Up Method
This is a variation of hook-up method 1 from above and is connected in the same way that you would bi-wire a speaker such as the RTi70. Two sets of speaker wires are run from the same set of terminals on the receiver/amp to the two separate sets of terminals on the speaker with the jumper plate removed.
While this is not bi-wiring by definition, it does accomplish a similar thing: it separates the midrange and tweeter signal from the subwoofer (bass) signal. This is why there is confusion over the term that is used to describe this hook-up method. I think that a new term needs to be coined for this hook-up method such as "powered tower bi-wiring."
To speculate a bit, I doubt that this hook-up method provides as much benefit as it does on a regular bi-wirable speaker for two reasons:
1. Even when using hook-up method 1, the amount of bass energy present in the speaker wire and crossover should be significantly less than in a passive speaker, because the subwoofer amplifier is only using the signal and then amplifying the driver itself (rather than using your receiver/amp's power to drive the subwoofer). I could be on this point, though.
2. In a passive speaker like the RTi70, the top posts would not see as low of a frequency as the powered towers do. Vice versa, the bottom terminals on the passive speaker would see a higher frequency than the powered tower. For example, the RT1000i crosses over between the midrange and subwoofer at 90Hz. Therefore, the top posts would see 90-20,000Hz. while the bottom posts would see 20-90Hz. In a passive speaker such as the RTi70, I would guess that the bottom posts would see a greater frequency range such as 20-200Hz. What this does is create a greater separation of the signal. More bass energy is removed from the midrange and tweeter signal (top terminals) and put with the bass signal (lower terminals). This should help to clean up the midrange and tweeter signal a bit more.
When you consider this, especially when coupled with point 1 above, it would seem that powered towers do not benefit as much from this hook-up method as do passive speakers. Regardless, this is just my theory, and I have no tested it.
Optimal Hook-up Method
The optimal hook-up method for powered towers is speaker-level connections. Specifically, I feel that there is some additional benefit to the "powered tower bi-wire" hook-up method. I can only see disadvantages to using the line-level connections unless you have a really old Pro-Logic system. I see two problems with line-level connections:
1. It would seem like this method would create some phasing (timing) problems between the midrange/tweeter signal and the subwoofer signal since you're sending them through different signal paths both in the receiver and the speaker. Also, you're likely sending them through different wire lengths as well.
2. This is probably the more important one. If you use line-level connections, the sonic signature (flavor of sound, if you will) will be different, because it has not been influenced (colored) by the amplifier. If speaker-level outputs are used, the signal that the subwoofer amplifier sees is going to have the same sonic signature as the rest of the signal seen by the midrange and tweeter. Still, the subwoofer amplifier will have a different sonic signature than the receiver, but this method will help to minimize the problem.
The only downside I see to speaker level connections for the powered tower subwoofer is that they will not be as clean as line-level signal. This is because the receiver has introduced its distortion/uncleanness into the signal. Overall I still feel this is the optimal hook-up method.
I hope you all found this discertation helpful, informative, and even entertaining.