Can you switch out drivers from another company???

El: Oops: got it backwards again! But sensitivity issues are still critical....
Re: the chicken soup...I leave that to my New York Jewish wife, who's version is incredible. I keep wanting to stretch it by adding more water, but Ellen's quick glare keeps me away from THAT pot!

Thanks for all your input. It sounds like it would be better to leave things be and eventually upgrade to new speakers.

Eldragon,
You're wrong. Doin' is much more fun!

Why not selling them, and buying what you are lookinf for? Shopping is alway fun.

Subaruguru...I suppose that making chicken soup can be turned into rocket science! :) In particular it isn't rocket science to know that a SMALLER capacitor (not a larger one) will raise the tweeter X/O frequency.

Tweeter substitution is a piece of cake compared with woofers, because enclosure considerations are minimal. (Spacing for phase reasons is about it). Just make sure that the new tweeter can go as low as the old one unless you change the X/O. Dispersion characteristics may change..better or worse is your opinion.

The suggestion by Subaruguru that modification of the X/O without replacing drivers might satisfy your needs is reasonable. However, bear in mind that someone else already "optimized" the crossover for these drivers. You might try substitution of same-value better-quality inductors and capacitors in the X/O. This ain't cheap. Again, do one speaker at a time so that you can evaluate the results.

Be careful. As Eldartford points out it is not rocket science but a lot goes into making a decent speaker. A lot of planning, listening and tweaking goes into making two drivers integrate well. I messed around with a two and half way system for about two and a half years a few year back (that is not why they call it two and a half way). Granted, I'm not the sharpest knife in the drawer, but I found out that it is not unusual to spend that kind of time on speakers. Guys who make good stuff most often measure their experience in decades. There are just a lot of things that go into it. The T/S parameters mentioned above are especially helpful in the base integration. Less so with the higher frequencies.

Anyway, check the frequency range of the new driver against the old. Basically, in simplest terms, the two drivers have to share certain frequencies and over-lap by at least a couple octaves. Whatever you do(as Subguru notes), do not ask the new tweeter to go lower than its comfort range. The natural/acoustic roll-off of the driver combines with the crossover's electrical roll-off so, for example, if you want a 12db roll-off you can't just use a 2nd order crossover and expect results.

As noted above, while it helps to initially check the driver's characteristics it is less than the whole story because the driver/cabinet combo has it's own unique characteristics and those are what really matter. You likely will not find any numbers on your proposed combo.
In addition, it is very hard to get matched drivers by some manufacturers and individual driver units can vary quite a bit.

Finally, after you tweaked them , even if it is good change, you are likely to find that you have decreased the value of your speakers upon resale if resale is an issue.

Nonetheless, if you have some spare time and want to learn something, go for it. It is just very unlikely, IMHO, that you are going to do something simple, like drop in a new set of drivers and make a change in the xo, and improve the sound much. More likely to be a long process and you want to be in a state of mind to enjoy it.

I remain,

You may get a radically different total power response (especially if sidewalls are close by) if polar radiation characteristics are different too. And then there's the basic resonant frequency, power handling vis a vis crossover point, not to mention the sensitivity and crossover-shaping major contouring. Not rocket science, but to many of our ears damned close! There's a lot to be potentially learned here, and in addition to basic design literature I'd suggest you invest under $100 in an analog Radio Shack SPL meter (NOT the digital), a decent tripod, and a test CD with 1/3 octave warbles, like the Stereophile one, et al, and run some simple curves from 1k up to 20k both in the nearfield (to sort out the two tweeters), and at listening distance (2-3m is good) to get overall room response. As you learn you'll be able to first be able to "pad" the tweeter(s) to approximate equal sensitivity, and then perhaps futz with the series cap into it for adjusting crossover point and power response. It can get a lot more complicated than that (bear with me, guys), but at least you'll develop an appreciation for the art, and indeed perhaps arrive at a sound you'll really like. Just be careful to not bring the new tweeter in at too low a crossover initially if it's power-handling isn't up to the Paradigm's. Starting at a higher cross-frequency (either by using a larger value cap or larger resistor) will protect it; then dial it in to the vicinity of the Paradigm's, tweaking for your preference. On the third hand you might indeed find that REDUCING the resistor pad on the Paradigm's a tiny amount is all you want to get a slightly brighter response (!), and save yourself all this effort in buying other tweeters! After all, tweeter level is a very subjective decision at the manufacturing/voicing level. If there's a 1.0 ohm resistor on the OE tweeter you might try a 3/4 ohm or 1/2 ohm and be done with it for $2! Have fun.

Suggestion...To minimize experimental cost, try doing one speaker, and then compare results with the other one. Assuming the nominal impedance of the two tweeters match, the crossover network will probably be OK except perhaps for the tweeter "padding" resistor, that adjusts its SPL to match the woofer. This is not rocket science, and your personal ears are the ultimate test equipment. Have fun.

At least TWO reasons why exchanging the tweet is tricky (but hardly inpossible): mechanical & electrical
Mechanical: does the new tweet fit, are the mechanical properties similar b/ween the two tweets?
Electrical: (more tricky) the Thiele-Small parametres must be very close (if not identical) -- otherwise you're changing the original design of your speakers / OR you'll have to change the cross-over to match the new tweet's characteristics (i.e. you're changing the original speaker design)...