This case illustrates the power of the architectural approach to the study of very complex systems and their [unintended] behavior.

It is well known that American politics have become highly polarized and rancorous in the last twenty or so years. I will show that an architectural approach can shed some plausible lights on the causes for this phenomenon.

Undenyably a root cause for this political rancor is the world wide ideological struggle between the millenary traditional religious belief systems and the "recent" religion of man.

However, besides this major WW phenomenon there are additional factors arising from the uniquely American political architecture & the development of computer technology (IT), that add force to the degradation of our political discourse.

As is well known, the American Constitution prescribes a total population census every ten years and a consequent redefinition of the House electoral districts to reflect the population changes of the last decade so that each voter is as nearly equipollent as possible.

The development of modern computers has impacted this decennial redistricting effort in a major way. In fact, in the last couple of decades it has been possible to push the art of district gerrymandering to levels never previously experienced.

Computers have no difficulty whatever to keep track of carloads of demographic & voter registration information & to perform very sophisticated computations to draw district boundaries so to maximize specific demographic or party affiliation parameters.

A well publicized example of such practices was the Texas gerrymander orchestrated by the former Speaker of the House, Tom Delay, in accord with the Black Caucus. The gerrymander resulted in creating a number of very oddly shaped districts with high concentrations of Black electors, while causing adjacent districts with White electors to have a reduced percentage of Democratic voting electors. The effect of this gerrymander was to increase the number of Black & Republican Congressmen elected out of Texas.

The next relationship to be considered is that occurring between such gerrymandered districts and Congress Members seniority. The gerrymandered districts because of their more homogeneous demographic or political makeup are more likely to keep reelecting their Congressman. Thus Members from such districts tend to acquire greater seniority than that of Members from more varied districts.

This seniority *propensity* can have several deleterious effects:

- Such Members can afford to be less responsive to their electors in
the pursuit of their
**ideological**policy preferences. - Members can be swayed by moneyed interests from outside their district

The bottom line is that the development of modern computer technology is causing an evolution of the House in the direction of greater power to the extreme ideologues of the Right & Left, extreme level of out-of-district money influence, very low level of responsiveness to the American public. Finally the increasing ideological polarization of the House results in a rancorous political interaction which tends not to solve the problems of the Nation.

It is obvious that gerrymandering is harmful to the health of the Republic thus it would be nice to reduce its occurrence. This in turn, is not easy to do by laws making it illegal. In fact, it is not obvious at first how one can provide a reliable objective test of gerrymandering to prove the crime.

A potential direction in which to look for such test is topology. Topology studies the property of geometric figures that do not change under non disruptive transformation. In plain English imagine the figure to be made of rubber and subject it to all kinds of stretching and pushing. From this point of view all closed boundary curves are equivalent to a circle since a circle can be deformed to match them. The circle has no bias at all so it should have a gerrymander index of zero. This is in fact the case if one looks at any two points on the circle and computes the factor:

- G = (r1 - r2)/r

Where r1 & r2 are the radius of the circle at the two points and r is the radius of the circle at any point. In the case of the circle r, r1, r2 are equal since all points on the circle are at a distance r from its center. Therefore the circle has a G = 0 or no gerrymandering at all.

For the case of arbitrary district boundary we can analogously say:

- G1 = (r1 - r)/r

- G = (G1 + G2+ G3 +....+Gn)/n

That is the gerrymander index would be the average of the boundary deformation
from the perfect (no gerrymander) circle. Laws prohibiting gerrymandering
could be easily written and **tested for compliance** if they would:
a) state some simple sampling rule for selecting the points on the boundary and
b) state the **maximum** G that would be tolerated, for example 20% or
30% or whatever the legislature would consider reasonable.

The last issue is how to find the center of the district. This is easily done
by creating a cutout with shape identical to the district and suspending
the cutout from two of its boundary points and noticing where a plumb line
crosses it. The two cross lines will intersect at the **center of gravity**
of the cutout. The corresponding point in the district is the district center.
If the cutout is a perfect scale replica of the district all the calculations
for G can be done on the cutout.

None of the currently emotional diatribes, such as toss the bastards out, will
solve this serious problems. Laws based on an objective & testable standard
are the **ONLY** way to put a stop to this practice which is destroying
our democracy.