[53] CLARK, Chief Judge, concurring specially:

This brief soliloquy is necessarily said, in my respectful view, because every other opinion goes farther than the Voting Rights Act intends. My brothers Gee and Higginbotham are at odds about the way the court should take to reach the same result. While their disagreement centers on the representative nature of the judicial office, the essence of their analyses of the impact of racial vote dilution in this judicial election process based on the nature of the office is similar — so similar that, if their opinions were expressly limited to the facts of the present case, I agree with both.

There is no disagreement that Section 2 of the Voting Rights Act, before its amendment, forbade any practice or procedure that abridged the right to vote because of race or color. All also agree that the legislative intent of the amendment was only to broaden the test for vote dilution from "intent" to "result."

The elements of Judge Gee's analysis are that, since section 2(b) defines vote dilution in terms of representatives, no vote dilution claim can be made in any election of a judicial officer because a judge can never be a representative — a conclusion he finds confirmed by the Supreme Court's refusal in Wells v. Edwards to apply one-man, one-vote standards to judicial election districts.

Judge Higginbotham rejects this analysis. He would base reversal on the premise that none of several elected trial judges who all function singularly in their work can be subject to the single-member redistricting claim made here. My concern is that the court's opinion, as now written, puts vote dilution attacks on (1) judicial elections which cannot be resolved by examining the nature of the office, and (2) "issue" elections (such as referenda on constitutional amendments and bond issue elections) beyond the reach of amended section 2.

Judge Gee starts with the observation that the words of section 2 expressly limit vote dilution to elections of representatives. I can readily agree section 2 does not apply to the elections challenged here. It involves only the election of persons and voter impact turns entirely on the nature of the judicial office. This brings section 2(b) into play. The inherent nature of the judicial function and, indeed, the constitutional limits of due process require that every judge be impartial between litigants and neutral as to claims presented. In the discharge of official duties, no judge can ever "represent" the electors in the jurisdiction served by the court. A vote for a judge differs from a vote for other types of officers. Whether the choice be for councilman, sheriff or governor, and whether it be based on whim or party or nonpartisan analysis of the individual candidate, votes for these types of officials are cast for those who will best express the wishes and views of their constituents. This cannot be so when a voter picks a judge. Legislators and executives are expected to represent. Voters must know judges cannot. The same principles control when a state provides for election rather than appointment of its judiciary. The choice seeks to assure the public that the judicial function will be kept accountable to the common sense of the electorate. It is expected that candidates who lack training or a reputation for honesty or sound intellect will not be elected. In like manner, those who are indolent, will not decide cases or decide erratically will not be re-elected. Overarching any considerations of voter motivation is the due process neutrality required in the conduct of the office. It does not permit the judge's responsiveness to the electorate's concept of common sense to become representation of the electorate. The State of Texas has a strong interest, and, indeed, a fundamental right to choose to have these judges elected in the manner provided here. Its choice does not violate amended section 2.

The difficulty I have with Judge Gee's analysis is that it has no limit. There are many types of elections which involve issues, not candidates, which surely ought to be subject to the vote dilution stricture of section 2 despite the absence of any question of representation. But merely noting the applicability to "issue" elections would not adequately define the reach of section 2. It is imperative, in my view, that a bright circle be drawn around judicial elections as well. Judge Gee's reasoning would expressly deny section 2(a) coverage to judicial elections in situations beyond today's facts, as he makes clear by overruling Chisom v. Edwards. Section 2(a) is an integral part of a remedial statute. It deserves to be interpreted so as to prevent racial vote abridgment even when it occurs in a judicial election. The phrase "as provided in subsection (b) of this section" which appears at the end of subsection (a) should be read as giving an example of proscribed vote dilution. It does not provide that section 2(b) establishes the only way the section can be violated.

It is clear to me that when a state continues to apply a voting procedure in a manner which now results in an abridgment of the right of a citizen to vote on account of race, that procedure is still condemned by amended section 2(a), just as it was before the amendment.

Nothing we say today should be taken as passing on a claim that a judicial election process in which judges are elected by fewer than all of the eligible voters within the jurisdictional area of the court on which the judge will serve has become a violation of section 2. Such elections involve districting of voters in a manner entirely unrelated to the representative nature of the judge's office.

[Thornburg v.] Gingles [ 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986)] tells me that whether the political process chosen by Texas for selecting its judges is equally open depends upon evaluation of past and present reality under a functional view of the process. There is nothing wrong with the state's choice to elect any number of a county's district judges county-wide. However, if the state has chosen to divide a single judicial jurisdiction into separate groups of electors, that choice could, with changes in demographics or other conditions come to raise real issues of racial gerrymander, gross diminution of voting strength, candidate slating ability or other violations of equal protection which have nothing to do with the due process concerns which control the execution of judicial duties, or with the manner in which the office of judge is carried out. Of course, I agree that Wells v. Edwards establishes that approximate numeric equality of voters between judicial districts is not required. However, we need not and should not decide now that judicial subdistricts which grow to have gross numeric or racial disparities in their make-up will always be free of possible section 2 problems. For this reason, I respectfully, but expressly, disagree with the majority's flat-out overruling of Chisom v. Edwards.

We are not confronted here with any claim of vote dilution resulting from long-established subdistricts alleged to have become racially invidious on a basis of intrajurisdictional voter distribution. This was the claim that was before this court in Chisom. The holding in Chisom reversed a dismissal on the pleadings. I agree that such a reversal was proper, even though I cannot agree with all said in Part I of Judge Higginbotham's concurrence or Judge Johnson's dissent because both deny vitality to section 2(b). Since we are writing en banc, I am free to disagree with the reason given for the result in Chisom — that section 2 applies to all judicial elections. I am of the opinion that it is equally wrong to say that section 2 covers all judicial elections as it is to say it covers none. However, if today's facts were the same as Chisom's, I would hold a claim that judicial subdistricts, once having no invidious purpose, but alleged, over time, to have come to abridge section 2 rights, must be factually developed and cannot be dismissed on pleadings alone.

If the issue were reached in today's case, I would also agree with Judge Higginbotham that the presence of multiple judicial posts on the ballots of plaintiffs here gives them no section 2 right to have single-judge subdistricts drawn. I would do so because I am not required to agree that the principle applies on any broader scale than the facts before us present. His function-of-the-office analysis is, to me, identical in concept to the majority view. The caveat I think must be added to both is that only when the area of jurisdiction of each of several jurist to be elected is coextensive with the area of residence of those that elect them, is each vote for a judge bound to be equal to every other vote that may be cast.

I would not agree with Judge Higginbotham that the single-judge, trial-court function of the judicial office is a critical factor. The analysis ought to be the same regardless of how the judge judges. When an appellate judge — who must function with other appellate judges to accomplish the judicial task — serves the same jurisdictional area as that which defines the electorate, section 2 does not allow a single member subdistricting remedy to be applied. This is so because no intradistrict or intrastate violation of section 2 is possible. The collegial nature of the appellate office in no way alters the compulsion for due process neutrality. When this neutrality is coupled with congruence of jurisdiction and electorate, they jointly assure equality in voting practices and procedures, negate representation and eliminate the possibility of vote dilution.

However, as with my partial agreement with Judge Gee's analysis, agreement with Judge Higginbotham should not be taken as controlling fact situations not before us here. The single-judge, trial-court functional analysis proceeds solely on what the judge does and the way he does it. These analyses change no basic principles. If the coincidence of voter residence and jurisdiction does not exist, the same possible vote dilution violations mentioned above, which have nothing to do with the function of the office being voted on, could occur. The importance of the policy embodied in section 2 compels me to say that these limits must be placed on what we write so that future courts will not cut short the intended reach of section 2. In my view, the majority view should be limited to the facts before us.

With the reservations expressed, I respectfully concur in reversing the judgment appealed from.

[68] HIGGINBOTHAM, Circuit Judge, with whom, POLITZ, KING and DAVIS, join, concurring in the judgment.[fn*] JOHNSON, Circuit Judge, concurs in Part 1. WIENER, Circuit Judge, specially concurs in Part 2 in addition to concurring in the majority opinion. [fn*] When this case was orally argued before and considered by the court, Judge Reavley was in active service. He participated in both the oral argument and the en banc conference. He took senior status, however, on August 1, 1990. Based on his understanding of the Supreme Court decision in United States v. American-Foreign Steamship Corp., 363 U.S. 685, 80 S.Ct. 1336,4 L.Ed.2d 1491 (1960), he considers himself ineligible to participate in the decision of this case, but he adheres to the views in this opinion. See Sawyer v. Butler, 881 F.2d 1273 (5th Cir. 1989) and Court Policy 21.C.

This is a voting rights suit challenging the election of district judges on a countywide basis in Texas. The suit was filed in a United States District Court by the League of United Latin American Citizens against the Attorney General of Texas, the Secretary of State, and other state officials, seeking a declaratory judgment that the at-large election of state district judges in nine targeted counties is illegal under Section 2 of the Voting Rights Act, 42 U.S.C. § 1973, and violative of the fourteenth and fifteenth amendments of the United States Constitution. Plaintiffs requested the district court to enjoin further elections and to impose a districting scheme that included single-member districts. Texas has 254 counties, but the suit attacked only Harris, Dallas, Tarrant, Bexar, Travis, Jefferson, Lubbock, Ector, and Midland Counties. These nine counties have more than one district judge elected county-wide, and elect 172 of the state's 390 district judges. As we will explain, the suit targets Texas law requiring election of a state district judge from a district no smaller than the county, the geographical area of its jurisdiction.

After a bench trial, the district court found violations of the Voting Rights Act in each of the nine counties, but rejected the constitutional arguments, finding that plaintiffs had failed to prove that the electoral system was instituted or maintained with discriminatory intent. On January 2, 1990, the district court enjoined defendants from

Calling, holding, supervising and certifying elections for state district judges in Harris, Dallas, Tarrant, Bexar, Travis, Jefferson, Lubbock, Hector and Midland Counties under the current at-large system with an order for interim relief.

The district court divided the nine counties into electoral sub-districts, tracing the districts of state representatives and the precinct lines of County Commissioners or Justices of the Peace. The district court's order affected 115 of the 172 district courts. The district court also ordered a non-partisan election for May 5, 1990, with any run-off to be held on June 2, 1990. We stayed the district court's order pending this appeal.

Defendants first argue that the Voting Rights Act as amended in 1982 has no application to the election of judges. This argument rests on the assertion that the use by Congress of the word "representatives" in Section 2(b), added by amendment in 1982 and popularly known as the Dole compromise, unambiguously excluded elected judges because elected judges are not representatives. This argument in its broadest form — Section 2 of the Act has no application to any judicial elections — was rejected by this court in Chisom v. Edwards, 839 F.2d 1056 (5th Cir.), cert. denied sub nom. Roemer v. Chisom, 488 U.S. 955, 109 S.Ct. 390, 102 L.Ed.2d 379 (1988). The panel opinion was unanimous. The petition for rehearing en banc was denied without a single member of the court requesting a poll. Relatedly, but with less sweep, defendants argue that Section 2(b) has no application to state district judges because such judges do their judging singly and not as part of a collegial body. Finally, defendants attack the findings below as well as the ordered remedy. In addition to quarrels with the sufficiency of proof that the votes of minorities were diluted, defendants argue that the findings are flawed by the erroneous legal conclusion that the contribution of partisan voting to election outcomes is not relevant.

We are unpersuaded that Chisom's decision regarding the election of appellate judges was incorrect, but are persuaded that Section 2(b) will not support this attack upon the countywide election of trial judges. Because we would decide the case on this ground we do not reach defendants' other contentions.

I. A.

We are pointed to no evidence of how the Framers' viewed elected judges. This is not surprising; judges were not elected at the time the Constitution was written and ratified. The thirteen original states employed various methods of judicial selection, seven using appointment by the legislature, five by governor and council, and one by governor and legislature. See Winters, Selection of Judges — an Historical Introduction, 44 Tex.L.Rev. 1081, 1082 (1966). Electing judges was a Jacksonian reform aimed at making judicial selection more democratic:

Popular sovereignty and popular control of public affairs through the elective system were hallmarks of the Jacksonian era, and, not surprisingly, the movement for popular election of judges dates from this period. Dissatisfaction with the judiciary was widespread among Jacksonians. It arose from several factors including a general disaffection with the legal profession, abuses in the judicial appointment systems, and a feeling, carried over from the Jeffersonian period, that the courts were basically undemocratic. Consequently, the abolition of tenure during good behavior and the adoption of the elective system were advocated as reform measures and were hailed as in accord with the egalitarian spirit of the times.

Note, The Equal Population Principle: Does It Apply to Elected Judges?, 47 Notre Dame L.Rev. 316, 317 (1971).

The first judicial elections took place as early as 1812 for Georgia lower court judges, Ga. Const. art. III, § 4 (1812), and in 1832 Mississippi adopted a completely elective judiciary. Miss. Const. art. IV, §§ 2, 11, 16 (1832). When it joined the Union, Texas ironically became the first new state to adopt the federal method of selecting judges, by executive appointment with confirmation by the state senate. Id.; Tex. Const. art. IV, Section 5 (1845). The wholesale change from appointed to elected judges can be marked by New York's adoption of judicial elections in 1846. N.Y. Const. art. VI, §§ 2, 4, 12, 14 (1846). All new states entering the union after that date, until the entrance of Alaska in 1958, used elections as their method of judicial selection, and Georgia, Maryland, Virginia, and Pennsylvania switched from appointment to election. Winters, Selection of Judges, 44 Tex.L.Rev. at 1082. In short, it is fair to conclude that electing judges was viewed as being more democratic and as a way of ensuring that judges remained sensitive to the concerns of the people.

It is vigorously argued that Section 2 of the Voting Rights Act has no application to judicial elections because judges are not representatives. The argument in its strongest form is that the word "representatives," found in Section 2(b), unambiguously excludes judges because judges have no constituents. The argument continues that there is no occasion for exploring legislative history because the inquiry ends with the plain words of the statute. While drawing the language of Section 2(b) from White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973), Congress substituted the word "representatives" for "legislators," at the least to insure it reached elected executive officials. This much defendants do not deny. Rather, they argue that although "representatives" may encompass executive officials, Congress intended that the term not encompass judges.

To be unambiguously inapplicable to judges, the word "representatives" must be certain of only one relevant meaning and that meaning must exclude judges. Defendants must concede, however, that at one level of generality judges are representatives. The history of electing judges and the political impulses behind that choice are powerful evidence of considered decisions to keep judges sensitive to the concerns of the people and responsive to their changing will. This reality belies the bold assertion that judges are in no sense representatives. The assertion that judges are not representatives actually masks a concern that judges should not be representatives. This is a choice left to the states. Convinced that direct accountability insures that judges represent the people in their judicial tasks, Texas has chosen to elect judges.

Judges are oath bound to obey the law and to make decisions in an impartial manner but that does not mean that they are in no practical sense representatives of the people. Yet, executive officials, who are considered representatives, are bound by the same oath. While judges are indeed far removed from the logrolling give and take of the legislative and even executive processes, the effort to assure "sensitivity" and "accountability" through elections is no more than an insistence that the judges represent the people in their task of deciding cases and expounding the law. State judges, wearing their common law hats, face decisions such as whether to adopt a comparative fault standard, and in doing so represent the people in a very real sense. At least at this level of generality judges are indisputably representatives of voters. Saying so in no way steps on the equally indisputable difference between judges and other representatives — that judges do not represent a specific constituency.

It is true that judges do not carry the views of a certain group of people into a larger governmental body, attempting to sway that body toward decisions favorable to their constituency. That is not the necessary role of a representative. We extoll the virtues of the jury in criminal cases — the jury is said to be the representatives of the people. Both judicial opinions and academic writings describe members of juries as representatives. See Spaziano v. Florida, 468 U.S. 447, 104 S.Ct. 3154, 3176, 82 L.Ed.2d 340 (1984); Gillers, Deciding Who Dies, 129 U.Pa.L.Rev. 1, 63-65 (1980); H. Kalven H. Zeisel, The American Jury 436 (1966). The examples can be multiplied, but the point is plain. The conclusion that the word "representative" has the singular meaning of legislator is nothing more than an effort to substitute judicial will for that of Congress. It is an undisguised effort by judges to claim for judges an exemption from the Voting Rights Act. This exercise of raw judicial power claims, for federal courts, power belonging to Congress and to the states. Texas has decided to elect its judges and Congress has decided to protect the rights of voters in those elections.

In sum, we cannot determine whether Section 2(b) of the Voting Rights Act applies to judicial elections by looking only to the word "representative." Rather, we must look to the context in which the word is used and legislative history, cautious as we must be over that enterprise. Exploration of this context requires that we determine whether in using the word representative in the 1982 amendments, Congress intended to withdraw the Act's existing coverage of judicial elections. That is, the freight the majority's use of representative must bear becomes enormous if, before the 1982 amendments, the Voting Rights Act reached judicial elections.

We therefore turn first to whether the Voting Rights Act covered judicial elections before 1982. We consider the 1982 amendments to the Act and review the legislative history of the amendments. We then turn to the question whether Congress was required to mention specifically the election of judges in the statute. The resolution of this question is informed by application of settled principles of federalism; we determine that the election of judges has no claim to the protections of federalism not shared by other institutions of state government. We next reject the argument that because the one-person, one-vote principle is inapplicable to the judiciary, racial vote-dilution claims under Section 2 must be inapplicable as well. Finally, we look at the interplay of Sections 2 and 5 to determine whether differences between the two sections preclude the application of Section 2 to judicial elections despite Section 5's coverage of those same elections, and conclude that they do not.

B.

Section 2, before the 1982 amendments, provided as follows:

§ 1973. Denial or abridgement of right to vote on account of race or color through voting qualifications or prerequisites.

No voting qualification or prerequisite to voting, or standard, practice, or procedure shall be imposed or applied by any State or political subdivision to deny or abridge the right of any citizen of the United States to vote on account of race or color, or in contravention of the guarantees set forth in section 1973b(f)(2) of this title.

42 U.S.C. § 1973 (1975).

Section 2 by its express terms reached state judicial elections. "Vote" or "voting" was defined as including "all action necessary to make a vote effective in any primary, special or general election ... with respect to candidates for public or party office and propositions for which votes are received in an election." 42 U.S.C. § 1973 l (c)(1). There was no mention of judges or the judiciary. There also was no mention of any other specific office. Judges are "candidates for public or party office" elected in a "primary, special, or general election." Congress intended to reach all types of elections, rather than to pick and choose. Indeed, even votes on propositions are within the purview of the Act. Section 14(c)(1), 42 U.S.C. § 1973 l (c)(1).

Defendants argue that the Act is silent as to judges, so it must be construed as not including judicial elections. They argue that, while judges in Texas are "candidates for public office," it is uncertain whether Congress, by providing a broad definition of "vote," also intended to create a private remedial cause of action of similar scope in Section 2. Congress expressly defined the term "vote" or "voting," however, and nothing suggests that Congress did not intend that definition to apply throughout the Act, including Section 2.

Congress intended that its 1965 Act provide protection coextensive with the Constitution. Justice Stewart reiterated this in City of Mobile v. Bolden:

[I]t is apparent that the language of section 2 no more than elaborates upon that of the Fifteenth Amendment, and the sparse legislative history of section 2 makes clear that it was intended to have an effect no different from that of the Fifteenth Amendment itself....

446 U.S. 55, 60-61, 100 S.Ct. 1490, 1495-1496, 64 L.Ed.2d 47 (1980). We reject the implicit suggestion that the protections of the Fifteenth Amendment do not extend to minorities whose right to vote in judicial elections is abridged. The Fifteenth Amendment applies to all elections, and Congress intended the Voting Rights Act of 1965 to apply to all elections.

By its terms the 1965 Act included judicial elections. Under defendants' argument then the word representative in Section 2(b) must bear the burden of being the sole means by which Congress in the 1982 amendments exempted judicial elections from the Act's coverage. The record is bereft of any hint that Congress's effort in 1982 to expand the Voting Rights Act carried a sub rosa withdrawal of coverage for state judicial elections.

C.

Congress amended Section 2 in 1982 in partial response to the Supreme Court's decision in City of Mobile v. Bolden, 446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980). Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 2758, 92 L.Ed.2d 25 (1986). Bolden held that in order to establish a violation under Section 2 of the Act a plaintiff must prove purposeful racial discrimination. Bolden, 446 U.S. at 66, 100 S.Ct. at 1499. Congress incorporated a "results test" into Section 2(a) to diminish the burden of proof necessary to prove a violation. Congress also added Section 2(b), which codified the legal standards enunciated in White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973). As amended in 1982, Section 2 now provides:

(a) No voting qualification or prerequisite to voting or standard, practice, or procedure shall be imposed or applied by any State or political subdivision in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color, or in contravention of the guarantees set forth in section 1973b(f)(2) of this title, as provided in subsection (b) of this section.

(b) A violation of subsection (a) of this section is established if, based on the totality of circumstances, it is shown that the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) of this section in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. The extent to which members of a protected class have been elected to office in the State or political subdivision is one circumstance which may be considered: Provided, that nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.

42 U.S.C. § 1973 (1982).

The plain language of Section 2(a) reaches judicial elections, using the same broad language as the 1965 Act, referring generally to "voting" and "vote," the definitions of which continued unchanged under the 1982 amendments. The legislative history of the 1982 amendments does not indicate that the terms "vote" or "voting" do not include judicial elections, or that "candidates for public office" does not include judges. While retaining the identical statutory reach, Congress added the word "results" as the measure of violation. The word representative does not appear in subsection (a).

Section 2(b) is a new section added in the 1982 amendments. Section 2(a) refers to "denial or abridgement of the right ... to vote on account of race or color ..., as provided in subsection (b) of this section." Section 2(a) anticipates that subsection (b) will define how a violation of subsection (a) can be established. Other than the previously discussed vague use of the word "representative," there is no reason to suppose that subsection (b), defining a type of proof sufficient under Section 2, was meant to withdraw all coverage from judicial elections. Before we turn to the legislative history of the 1982 amendments for evidence of intent to exclude judicial elections from coverage, we pause to emphasize that the exercise is itself not necessary. A straightforward reading of both Sections 2(a) and 2(b) leaves little doubt but that Section 2(a)'s broad reach was never intended to be limited by use of the word representative in the explanation in Section 2(b) of how a violation might be shown.

Congress used the word "candidates" interchangeably with "representatives" in the legislative history. There was no indication that "representatives" was intended to have a limited meaning, applying only to legislative and executive officials, but not to elected members of the judiciary. Even Senator Dole, who proposed the language of compromise in Section 2, stated

Citizens of all races are entitled to have an equal chance of electing candidates of their choice, but if they are fairly afforded that opportunity, and lose, the law should offer no redress.

S.Rep. No. 417, 97th Cong., 2d Sess. 193 (Additional Views of Senator Robert Dole), reprinted in 1982 U.S. Code Cong. Admin.News 177, 364 (emphasis added), and

[T]he standard is whether the political processes are equally "open" in that members of a protected class have the same opportunity as others to participate in the political process and to elect candidates of their choice.

Id. (emphasis added).

In the one place where the judiciary is specifically mentioned in the legislative history of the 1982 amendments, the report of the subcommittee on the Constitution states that the term "`political subdivision' encompasses all governmental units, including city and county councils, school boards, judicial districts, utility districts, as well as state legislatures." Report of the Subcommittee on the Constitution of the Committee of the Judiciary, S.Rep. 417, 97th Cong., 2d Sess., reprinted in 1982 U.S. Code Cong. Admin.News 177, 323 (emphasis added). Of course, a brief statement in a subcommittee report opposing the amendments is not much. Nonetheless, the proponents of the changes to the Act did not contest this description, although they would have had incentive to do so to alleviate any fears of such extended coverage if such a broad scope of applicability were not intended.

The Senate and House hearings regarding the 1982 amendments contain various references to judicial elections, primarily in the context of statistics presented to Congress indicating the progress made by minorities under the Act up to that date. The charts indicated when minorities were elected to office, and included judicial election results. See Extension of the Voting Rights Act: Hearings on H.R. 1407, H.R. 1731, H.R. 3112, H.R. 3198, H.R. 3473 and H.R. 3498 Before the Subcomm. on Civil and Constitutional Rights of the House Comm. on the Judiciary, 97th Cong. 1st Sess. 38, 193, 239, 280, 502, 574, 804, 937, 1182, 1188, 1515, 1528, 1535, 1745, 1839, 2647 (1981); Voting Rights Act: Hearings on S. 53, S. 1761, S. 1975, S. 1992, and H.R. 3112 Before the Subcomm. on the Constitution of the Senate Comm. on the Judiciary, 97th Cong. 2d Sess. 669, 748, 788-89 (1982).

To summarize, the relevant legislative history concerning the 1982 amendments suggests that Section 2(b) was intended to reach all elections, including judicial elections. There is no hint that Congress intended to withdraw coverage.

But, it is argued, even if other aspects of Voting Rights law do apply to judicial elections, vote-dilution claims should not, because these claims are a new and fundamentally different ground for relief under amended Section 2 and because anti-dilution remedies are particularly intrusive on the judiciary. Therefore, the argument continues, had Congress intended the Act to apply to judicial elections, it should have said so explicitly, which it did not. We reject this argument that Congress singled out both judicial elections and dilution claims for distinct treatment. In plain language it argues that Congress affirmatively turned its head away from the dilution of minority votes in judicial elections.

The first flaw in this argument is that vote-dilution claims were not newly authorized by amended Section 2. There were many vote dilution cases before 1982. The statutory prohibition of vote dilution by the Voting Rights Act is as old as the Act itself. It was first raised as early as 1965, the year of the Act's inception, when the Supreme Court observed that

[i]t might well be that, designedly or otherwise, a multi-member constituency apportionment scheme, under the circumstances of a particular case, would operate to minimize or cancel out the voting strength of racial or political elements of the voting population. When this is demonstrated it will be time enough to consider whether the system still passes constitutional muster.

Fortson v. Dorsey, 379 U.S. 433, 439, 85 S.Ct. 498, 501, 13 L.Ed.2d 401 (1965). Vote-dilution claims were considered in Burns v. Richardson, 384 U.S. 73, 86 S.Ct. 1286, 16 L.Ed.2d 376 (1966), and Whitcomb v. Chavis, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363 (1971), where the plaintiffs were unsuccessful, and in White v. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973), and Zimmer v. McKeithen, 485 F.2d 1297 (5th Cir. 1973) (en banc), aff'd sub nom. East Carroll Parish School Board v. Marshall, 424 U.S. 636, 96 S.Ct. 1083, 47 L.Ed.2d 296 (1976), where the plaintiffs prevailed. These cases were decided under the results test. Finally Mobile v. Bolden, 446 U.S. 55, 100 S.Ct. 1490, 64 L.Ed.2d 47 (1980), where the Supreme Court articulated the intent standard, was a dilution case. The 1965 Act, therefore, was considered to prohibit vote dilution as well as more straightforward denials of the right to vote. By its terms the act covered judicial elections. The 1982 amendments simply made it clear that results and not intent were the basis for finding a violation. However difficult in application the results test may have proved to be, the amendments to Section 2 did not themselves create a vote-dilution claim. To the contrary, the dilution of the voting strength of minorities was the accepted premise of the debate. Indeed Zimmer v. McKeithen, 485 F.2d 1297 (5th Cir. 1973) (en banc), the source of the "senate factors" that became part of the congressionally required inquiry was a dilution case.

Much of the legislative history of the 1982 amendments indicates that Congress intended to return to pre- Bolden standards and was not otherwise reaching for a new and more intrusive private cause of action. As we will explain, at least Senator Hatch feared the language of the 1982 amendment would be much more intrusive, expressing concern that its uncertainty would lead to proportional representation. His fear was fueled by the restoration of the results test, however, not dilution theory, which had been part of the voting rights law for at least seventeen years.

The principal focus of the House debates centered on Section 5, but the Senate debates were centered on the meaning of the Section 2 amendments. Nonetheless, there was some discussion in the House, and at least some witnesses argued that "the amended Section 2 ... would restore to black Southerners the right to challenge alleged discriminatory election schemes which were developing before Mobile, [and that] notwithstanding the Court's claim to the contrary in Mobile, the intent test first became a constitutional standard in 1976 with Washington v. Davis, an employment case." Boyd Markman, The 1982 Amendments to the Voting Rights Act: A Legislative History, 40 Wn. Lee L.Rev. 1347, 1366 (citing comments by James Blacksher and David Walbert). Congressman Sensenbrenner argued that the Rodino amendment to Section 2 was necessary in order to clarify the standard of proof required in order to establish violations of the Act. 127 Cong.Rec. H6850 (daily ed. Oct. 1981) at H6983.

In the Senate Report on the Amendments the purpose of the bill was stated as

designed to make clear that proof of discriminatory intent is not required to establish a violation of Section 2. It thereby restores the legal standards based upon the controlling Supreme Court precedents, which applied in voting discrimination claims prior to the litigation involved in Mobile v. Bolden. The amendment also adds a new subsection to Section 2 which delineates the legal standards under the results test by codifying the leading pre-Bolden vote dilution case, White v. Regester.

S.Rep. 417, 97th Cong., 2d Sess., reprinted in 1982 U.S. Code Cong. Admin.News at 179 (emphasis added).

Senator Hatch opposed the change, arguing that it "would redefine the concept of `discrimination' and would `transform the Fifteenth Amendment and the Voting Rights Act from provisions designed to ensure equal access and equal opportunity in the electoral process to those designed to ensure equal outcome and equal success.'" Boyd, Voting Rights Act Amendments, 40 Wn. Lee L.Rev. at 1389 (quoting Hearings on the Voting Rights Act Before the Senate Subcommittee on the Constitution of the Committee on the Judiciary, 97th Cong., 2d Sess. 3 (1982)). But, Senator Mathias, a proponent of the bill, argued that

[t]he House amendment is needed to clarify the burden of proof in voting discrimination cases and to remove the uncertainty caused by the failure of the Supreme Court to articulate a clear standard in the City of Mobile v. Bolden.... We are not trying to overrule the Court. The Court seems to be in some error about what the legislative intent was.... Prior to Bolden, a violation in voting discrimination cases [could] be shown by reference to a variety of factors that, when taken together, added up to a finding of illegal discrimination. But in Bolden, the plurality appears to have abandoned this totality of circumstances approach and to have replaced it with a requirement of specific evidence of intent ... this is a requirement of a smoking gun, and I think it becomes a crippling blow to the overall effectiveness of the Act.

Hearings on the Voting Rights Act Before the Senate Subcommittee on the Constitution of the Committee on the Judiciary, 97th Cong., 2d Sess. 3, 199 (1982).

Senator Hatch persisted that the results test represented a new test, but supporters of the bill took issue with this view. Laughlin McDonald of the ACLU argued that "[p]rior to Mobile, it was understood by lawyers trying these cases and by the judges who were hearing them that a violation of voting rights could be made out upon proof of a bad purpose or effect ... Mobile had a dramatic effect on our cases." Id. at 369. Defenders of the amendment assumed that the results test represented a restatement of the law prior to Bolden.

Critics of the results test argued that even if the lower federal courts had adopted a results test in their pre- Bolden interpretation of Section 2, the original intent of Congress had been the establishment of a test in Section 2 using the traditional standard of intent or purpose. Boyd, Voting Rights Act Amendments, 40 Wn. Lee L.Rev. at 1405 (citing Appendix to Additional Views by Senator Hatch, S.Rep. No. 417, 97th Cong., 2d Sess. 36 (1982)). Proponents responded by arguing that there was no evidence that Congress meant an intent test to apply. The Senate Report of the Committee on the Judiciary adopted this view, citing Attorney General Katzenbach's testimony during the hearings on the Voting Rights Act of 1965 to the effect that "Section 2 would ban `any kind of practice ... if its purpose or effect was to deny or abridge the right to vote on account of race or color." S.Rep. 417, 97th Cong., 2d Sess., reprinted in 1982 U.S. Code Cong. Admin.News at 194 (citing Hearings on S. 1564 before the Committee on the Judiciary, 89th Cong., 1st Sess., 191 (1965)).

This legislative history generally indicates an intent to retain pre- Bolden standards rather than create a more intrusive new cause of action. We have insisted in other contexts that Congress clearly state its intent to supplant traditional state prerogatives. Judicial insistence upon clear statement is an important interpretative tool vindicating concern for separation of powers and federalism. See Atascadero State Hospital v. Scanlon, 473 U.S. 234, 105 S.Ct. 3142, 87 L.Ed.2d 171 (1985); Pennhurst State School and Hospital v. Halderman, 465 U.S. 89, 104 S.Ct. 900, 79 L.Ed.2d 67 (1984) ( Pennhurst II). This insistence upon "an unequivocal expression of congressional intent," Pennhurst II, 465 U.S. at 99, 104 S.Ct. at 907, is based upon the fundamental nature of the interests at stake, Atascadero, 105 S.Ct. at 3147 ("The `constitutionally mandated balance of power' between the states and the Federal Government was adopted by the Framers to ensure the protection of `our fundamental liberties.'") (quoting Garcia v. San Antonio Metropolitan Transit Authority, 469 U.S. 528, 572, 105 S.Ct. 1005, 1028 (Powell, J., dissenting)). These mighty principles do not carry us very far here. Congress has clearly expressed the Act's application to the states, and has clearly expressed its intent that violations of the Act be determined by a results test rather than an intent standard. By these actions, the Act, with all of its intrusive effect, has been made to apply to the states. Significantly, the "results tests" did apply to all elections including judicial elections until the 1980 decision of Mobile v. Bolden, supra. Thus, intrusive as it is, the Act, including the anti-dilution provisions, applied to judges before the 1982 amendment. The suggestion that anti-dilution and results tests were introduced by the 1982 amendments is wrong.

The majority's argument is by necessity a demand for the exemption of judicial elections from the Act as a whole. We cannot recognize this broad exemption. Section 5, commonly seen as the most far reaching of the Voting Act provisions, see South Carolina v. Katzenbach, 383 U.S. 301, 358-62, 86 S.Ct. 803, 833-35, 15 L.Ed.2d 769 (1966) (Black, J., dissenting), has allowed no escape for elected state judiciaries. Haith v. Martin, 618 F. Supp. 410 (E.D.N.C. 1985), aff'd mem., 477 U.S. 901, 106 S.Ct. 3268, 91 L.Ed.2d 559 (1986). As an inferior court we are bound by the holding of the Supreme Court that judicial elections are covered by Section 5 of the Act, a result explicitly urged by then Solicitor General Charles Fried and by then head of the Civil Rights Division, Assistant Attorney General William Bradford Reynolds. The same officials argued in Chisom that amended Section 2 of the Act is equally applicable, as does the present administration.

D.

Finally, it is argued that an elected state judiciary is somehow free of the anti-dilution prohibitions of the Voting Rights Act but remains subject to its other strictures. The argument has two premises: First, because the pre- Bolden anti-dilution cases did not involve judicial elections, Section 2's prohibition against vote dilution does not extend to judicial elections; second, because the one-person, one-vote principle has been held not to apply to judicial elections, vote-dilution claims under Section 2 do not apply either.

The first premise is obviously flawed. Nothing in the pre- Bolden cases suggests that the prohibition against vote dilution does not apply to judicial elections. That those cases involved elections of officials other than judges was happenstance; cases involving judicial elections simply had not yet come up. Furthermore, the statutory language cannot be parsed to read that judicial elections are not subject to dilution claims, but are subject to the remaining strictures of Section 2. This is so even if representative is found to mean elected members of the legislative and executive branches but not the judicial branches of state government. Further, concluding that Section 2 does not apply would create the anomaly that Section 5, conceded to reach elected judges, and Section 2 use identical language to define their reach. Section 2 either applies in its entirety or not at all and defendants' efforts to soften the full force of their extraordinary contention must fail.

The second premise — that because the one-person, one-vote principle does not apply to judicial elections, the vote-dilution prohibition does not either — must also fail. The prohibition of geographical discrimination in voting expressed in Baker v. Carr, 369 U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962), and Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964), commonly referred to as the one-person, one-vote principle, was held not to apply to the apportionment of state judiciaries in Wells v. Edwards, 347 F. Supp. 453 (M.D.La. 1972) (3-judge court), aff'd mem., 409 U.S. 1095, 93 S.Ct. 904, 34 L.Ed.2d 679 (1973) (three justices, dissenting). It is argued that vote-dilution principles cannot be applied to an elected judiciary because the one-man, one-vote principle does not apply, and without requiring equal apportionment there is no benchmark for concluding that there is vote dilution. This argument rests upon the equating of racial and non-racial acts by the state that deny voting strength. Yet they measure equality on quite different planes. One is facially neutral in the matter of race; indeed compliance may adversely affect black voting power. The other rests on core concerns of the Civil War amendments — submerging of minority voting strength by the combined force of election methods and bigotry. In the more concrete terms of this case, that the state has chosen to allot thirty-some judges to Dallas County and only one to another county is not relevant. Submerging votes of protected minorities by a cohesive white majority is relevant.

It is perverse now to reason that because the elections of state judges are free of the Reynolds' command of numerical equality, an elected judiciary is a fortiori free from the racial equality commands of the Civil War Amendments and the Voting Rights Act. It is perverse because even the defenders of the "political thicket" doctrine at all times maintained that the courts must hold to the core values of the Civil War Amendments. For example, Justice Frankfurter, in his famous dissent to the Court's entry into the political thicket in Baker v. Carr admitted, joined by Justice Harlan, that "explicit and clear constitutional imperatives guided judicial intervention in state government on the issue of black disenfranchisement." Baker v. Carr, 369 U.S. 186, 285-86, 82 S.Ct. 691, 747-48, 7 L.Ed.2d 663 (1962) (Frankfurter, J., dissenting).

The courts have struggled to develop a measure of dilution stemming from the combination of racial voting patterns and state election practices. Gingles itself was the first detailing of that enterprise by the Supreme Court. At earlier times, various justices have referred to our efforts to do so in Zimmer v. McKeithen as amorphous. But, this difficulty has nothing to do with the inapplicability of the command of numerical equality, nor is its difficulty peculiar to judicial elections. We remind that the effort in this case to measure the submerging of black and Hispanic voting power begins with a system that is numerically perfect — county-wide elections.

We are pointed to several lower court opinions stating that judges are not "representatives." These cases held that the prohibition against geographical discrimination does not reach judicial elections. The argument is that because many of these courts held that judges are not "representatives," Congress must have meant a similar exclusion in its use of the word. We disagree. Words come to their full meaning in context. This argument of incorporated definition is unsupported by a trace of legislative history and is no more than an assertion. Indeed we showed above that Congress meant "representative" to include judges for the purposes of the Voting Rights Act. The Reynolds principle is race neutral, different, as we observed, from the race-based focus of the Voting Rights Act. However problematic locating the principle of one-person, one-vote in the fourteenth amendment may be, race-based concerns are at its core. Nothing in policy or logic suggests that declining to extend the Reynolds principle to judicial elections carries any sway in freeing judicial elections from race-focused concerns.

Wells was distinguished from cases challenging election practices in Lefkovits v. State Board of Elections, 400 F. Supp. 1005 (N.D.Ill. 1975) (3-judge court), aff'd mem., 424 U.S. 901, 96 S.Ct. 1092, 47 L.Ed.2d 306 (1976), where the court stated:

[W]hen a judge is to be elected or retained, regardless of the scheme of apportionment, the equal protection clause requires that every qualified elector be given an equal opportunity to vote and have his vote counted.

Id. at 1012. This was the precise point made by Solicitor General Fried in his successful argument to the Supreme Court that it should summarily affirm Haith v. Martin.

In Haith the district court held that judicial elections are covered by Section 5 and the preclearance requirements of the Act. The district court found, using an analysis similar to that used by this circuit in Voter Information Project v. Baton Rouge, 612 F.2d 208 (5th Cir. 1980), that although the one-person, one-vote principle may not apply to judicial elections, claims with respect to the Voting Rights Act do not deal with numerical apportionment, but with discrimination. The court held that "the Act applies to all voting without any limitation as to who, or what, is the object of the vote." 618 F. Supp. at 413. In short, Haith rejects the suggestion that inapplicability of the Reynolds principle is any barrier to the application of the Voting Rights Act. We are bound by Haith, and the relevance of that bind turns on whether Section 5, dealt with in Haith, and Section 2 are coextensive in their application to the judicial elections. We turn now to that question.

E.

Defendants have raised no compelling reason to distinguish between Section 5 and Section 2 with respect to their applicability to judicial elections. To distinguish the Sections would lead to the incongruous result that if a jurisdiction had a discriminatory voting procedure in place with respect to judicial elections it could not be challenged, but if the state sought to introduce that very procedure as a change from existing procedures, it would be subject to Section 5 preclearance and could not be implemented. Sections 2 and 5 operate in tandem, with Section 2 prohibiting the continued use of discriminatory practices, and Section 5 preventing the imposition of new discriminatory practices to replace those condemned in those areas subject to preclearance. Section 5 contains language defining its scope that is almost identical to the language in Section 2: "any voting qualification or prerequisite to voting, or standard, practice, or procedure with respect to voting. . . ."

There are important differences in the two sections, however. Section 5 requires preclearance of any new voting practices and procedures, and, in determining whether or not a new practice is entitled to preclearance, only the effect of the new practice is considered. City of Lockhart v. United States, 460 U.S. 125, 103 S.Ct. 998, 74 L.Ed.2d 863 (1983); Beer v. United States, 425 U.S. 130, 96 S.Ct. 1357, 47 L.Ed.2d 629 (1976). This has been described as a retrogression test, with preclearance denied only if the new practice has a retrogressive effect, rather than a results test, for the effects of the existing system on minorities are not considered. Thus in Section 2 the entire scheme of voting practices and procedures is considered to see whether it results in less than an equal opportunity to participate in the political process, whereas under Section 5 only the effects of new practices and procedures are considered. Section 2 is, therefore, arguably more intrusive than Section 5, for Section 5 only regulates whether or not changes may be implemented, whereas Section 2, if a violation is found, can lead to the dismantling of an entire system of voting practices that may have been in place for many years. This is a distinction between the two sections, but our question must be whether the difference means that Section 5 applies to judicial elections, but Section 2 does not. There appears to be no relevant reason why judicial elections are so different from legislative or executive elections that both sections should apply to one and not the other.

The Voting Rights Act plainly covered judicial elections before the 1982 amendments. It is equally plain that there is little evidence that Congress intended any retrenchment by its 1982 amendments. In sum, defendants are left with the unconvincing argument that the changes of the 1982 amendments were fundamental in ways unique to judicial elections. Certainly, the Voting Rights Act intrudes heavily into state matters but it is no more specifically intrusive in judicial elections than in any others. We would hold that Section 2 of the Voting Rights Act applies to judicial elections.

II.

We now turn to the quarrel with the county-wide election of Texas trial judges. The Voting Rights Act does not purport to change the choices by a state of the duties and means for their discharge it gives to a particular office it chooses to create. Rather, the Act accepts these state creatures but patrols for impermissible vote dilution of minority voting power caused by the features of the election process in combination with racially molded voting patterns in any election of such officials. The statute, however, gives no right to choose how the combination will be broken. It is important, then, that we keep in mind that the analysis of Thornburg v. Gingles is relevant only to an inquiry into whether an at-large election impermissibly dilutes minority voting strength; it is not a way of assessing every claimed vote dilution.

Texas has structured its government such that it wields judicial power at the trial level through trial judges acting separately, with a coterminous or linked electoral and jurisdictional base, each exercising the sum of judicial power at that level, and all with review by courts acting collegially. We are persuaded that, for purposes of the Voting Rights Act, because the fact and appearance of independence and fairness are so central to the judicial task, a state may structure its judicial offices to assure their presence when the means chosen are undeniably directly tailored to the objective. The choice of means by Texas here — tying elective base and jurisdiction — defines the very manner by which Texas' judicial services are delivered at the trial court level. These means define the office. Nothing in the Voting Rights Act grants federal courts the power to tamper with these choices. It requires no narrow reading to conclude that the statute does not by its terms purport to do more. Stated in traditional Fourteenth-Amendment terms, there is compelling necessity sufficient to overcome the strict scrutiny of state acts impinging upon a fundamental interest. We would not lightly suppose that the Voting Rights Act reached further than the Civil Rights Amendments except for dispensing with the requirement of purposeful violation.

It follows that inquiry into the Section 2 claims proceeds by accepting that trial judges are officials exercising the full authority of their positions alone whose full authority has its source in the electors from a district coterminous with their jurisdiction. There can be no dilution of votes for a single judge because each judge holds a complete judicial office. This feature of the trial judge will alone decide this case but, as we will explain, we need not rest only on this proposition. Rather, that the trial judges act singly is also integral to the linking of jurisdiction and elective base.

A.

The district courts are the primary trial courts in Texas. Indeed, the constitution of the Republic of Texas provided:

The Republic of Texas shall be divided into convenient judicial districts, not less than three, nor more than eight. There shall be appointed for each district a judge, who shall reside in the same, and hold the courts at such times and places as Congress may by law direct.

Guittard, Court Reform, Texas Style, 21 Sw. L.J. 451, 456 (1967). The first state constitution, adopted in 1845, contained essentially the same provision in article IV, section 6. This provision was amended in 1850 to allow for the election of district judges by the people, but the subsequent constitution of 1861 provided that district judges were to be appointed. Tex. Const. art. V, § 7, interpretive commentary (1876, amended 1985). Texas constitutions adopted since 1861, including the current constitution, which was adopted in 1876, have provided for elected district judges.

All the constitutions have provided that the district courts are to be held by district judges chosen from defined districts, following the pattern of the Constitution of the Republic of Texas. Although in the Constitution of the Republic of Texas the number of district courts was limited to not more than eight, subsequent constitutions have left the number of courts to the legislature. All Texas constitutions, including the current one, before it was amended in 1985, suggested that each district would be served by only one judge. See Tex. Const. art. V, § 7 (1876, amended 1985) ("[f]or each district there shall be elected ... a Judge...."). A one judge per district system, however, presupposes districts of substantially equal population. Guittard, supra at 456. Thus, with the growth of the population in certain counties it became necessary for the legislature to make adjustments.

The system challenged in this case was set up according to this pattern. See Tex. Gov't Code §§ 24.001-.954 (Vernon 1988 Supp. 1990). With the exception of the 72nd district, each challenged judicial district in the nine targeted counties is coextensive with one county. The 72nd district is composed of two counties. Id. § 24.174 (Vernon 1988). Since 1907 district judges have been elected county-wide. In 1985, however, a section was added to article V of the 1876 Constitution which specifically allows the creation of judicial districts smaller than a county. Tex. Const. art. V, § 7a(i) (1985). A majority of the voters in the county must authorize the division. Id. This power has yet to be exercised.

The district courts in multi-district counties were unified for certain administrative purposes in 1939 through the passage of the Special Practice Act, which is now, for the most part, found in Tex.R.Civ.P. 330(e)(i). Guittard, supra at 457-58. The relevant parts of the Special Practice Act essentially provide that cases can be freely transferred between judges and that any judge can work on any part of a case including preliminary matters. Also, "[a]ny judgment rendered or action taken by any judge in any of said courts in the county shall be valid and binding." Tex.R.Civ.P. 330(h).

The Administrative Judicial Act, originally passed in 1927 and subsequently amended on several occasions, divides Texas into nine administrative regions, each with a presiding judge appointed by the governor with the advice and consent of the senate. See Tex. Gov't Code §§ 74.005, .042 (Vernon 1988). The "presiding administrative judge is the key administrative officer in the Texas judicial system." Guittard, supra at 459. He is empowered to assign judges as necessary within his region. Id. §§ 74.052-056 (Vernon 1988 Supp. 1990); see also Judicial Administration Rule 8 (Vernon 1988 Supp. 1990). He is required to call two meetings of all judges in his administrative region each year and any other meetings as necessary. Tex. Gov't Code § 74.048(a) (Vernon 1988); Judicial Administration Rule 4 (Vernon 1988 Supp. 1990). This conference is for "consultation and counseling concerning the state of the civil and criminal business" and is empowered to promulgate administrative rules, rules governing the order of trials and county-wide recordkeeping, and other rules deemed necessary. Tex. Gov't Code § 74.048(b)-(c) (Vernon 1988).

The local administrative judge is elected by a majority vote of all the judges in the county, including both district and statutory judges. Id. § 74.091 (Vernon 1988 Supp. 1990). His duties on the county level are similar to those of the presiding administrative judge. See id. § 74.092. The local administrative judge has the power to assign judges within his county. Id. § 74.094. Under the leadership of the local administrative judge, the district and statutory judges in each county are directed to adopt local rules of administration. Id. § 74.093. These rules must provide for, among other things, the "assignment, docketing, transfer, and hearing of all cases" and "fair and equitable division of caseloads." Id. § 74.094(b); see also Judicial Administration Rule 9(b) (Vernon 1988 Supp. 1990). All local rules, of course, must be consistent with state and regional rules. Judicial Administration Rule 10 (Vernon 1988). In this regard, the present Chief Justice of Texas testified at trial that the only collegial decision-making by district judges in a county is in the handling of some administrative matters.

B.

A distinction was drawn between multi-member and single-member structures in Butts v. City of New York, 779 F.2d 141 (2d Cir. 1985), cert. denied, 478 U.S. 1021, 106 S.Ct. 3335, 92 L.Ed.2d 740 (1986). In that case the plaintiffs contested a primary run-off law, contending that it violated the Equal Protection Clause and the Voting Rights Act. The Second Circuit noted that one of the ways that a class of citizens may have less opportunity to participate is when there are electoral arrangements that diminish a class's opportunity to elect representatives in proportion to its numbers. The court distinguished, however, between multi-member bodies, where at-large elections may produce this result, and elections for single-member offices. Butts, 779 F.2d at 148. The court found that the Supreme Court had made this distinction implicit in City of Port Arthur v. United States, 459 U.S. 159, 103 S.Ct. 530, 74 L.Ed.2d 334 (1982), where the Supreme Court struck down a run-off requirement for seats on a multi-member city council, but did not mention the run-off requirement for mayor. The Eleventh Circuit followed Butts in United States v. Dallas County, Ala., 850 F.2d 1430 (11th Cir. 1988), in holding that "the at-large election of the probate judge is permissible under the Voting Rights Act with respect to the judicial aspects of that office." Id. at 1432 n. 1.

The positions at issue in Butts and Dallas County, and the position not considered in Port Arthur, were what can be viewed as traditional single member offices, i.e., mayor, city council president, single probate judge, or comptroller. There was only one of each office in a given geographical area, and no problem with overlapping jurisdictions. Here, there are many judges with overlapping jurisdictions. Nonetheless, each acts alone in wielding judicial power, and once cases are assigned there is no overlap in decision-making.

Indeed there are special courts created within some judicial districts that emphasize the single-member nature of the offices, for not all of the judges handle the same type of work. Some are courts of general jurisdiction, but some judges are elected specifically to handle juvenile cases, or family law cases, or criminal cases. To that extent they are separate offices, just as county treasurer and sheriff are different positions. Of course, many of the judges do handle the same type of cases and the cases are assigned to any of these judges within a given geographical jurisdiction. There are many of them within a geographical area, and the plaintiffs would find this dispositive. A United States district court in Alabama has held that Alabama trial courts similar to the Texas courts are multi-member positions. Southern Christian Leadership Conf. v. Siegelman, 714 F. Supp. 511 (M.D.Ala. 1989). The court considered Dallas County and Butts, but concluded that

Although neither court expressly defined the term "single-member office," it is clear to this court that the phrase, as used in those cases, refers to a situation where under no circumstances will there ever be more than one such position in a particular geographic voting area.

Siegelman, 714 F. Supp. at 518.

The court found that exclusive authority alone does not define single-member official. Id. We disagree with this view of multi-member versus single-member office and agree with the argument made by defendants in Siegelman that

the hallmark of a single member office, as [the Butts and Dillard v. Crenshaw County, 831 F.2d 246 (11th Cir. 1987)] courts use the term, is not the fact that the office is traditionally held by only one individual but, more importantly, the fact that the full authority of that office is exercised exclusively by one individual.

714 F. Supp. at 518.

Viewing district judges as members of a multi-member body is flawed in concept. Before any suits are filed, before any cases are assigned, there is a group of judges with concurrent jurisdiction, and plaintiffs maintain that this group should have minority members, so that minorities' views and concerns are considered by the judges who decide important issues in their lives. The problem is that once a case is assigned, it is decided by only one judge. The other judges have absolutely no say over the disposition of that case, and no influence over the deciding judge. One commentator has described the Texas system as a "one-judge, one court organization at the trial level, with rigid jurisdictional lines and with each judge largely independent of any supervisory control, except by way of appellate review." Guittard, Court Reform Texas Style, 21 Sw. L.J. at 455.

C.

It is implicit in Gingles that the effect of election practices must be considered after taking the underlying definition of the offices of state government as given. Even the sharply divided Gingles Court agreed that its inquiries were only into the legality of at-large methods of electing representatives to a larger governing body. Section 2 does not grant federal courts the authority to disregard the states' basic arrangements. We would not rest on inference to support such a grant of authority. It would run counter to fundamental concepts of federalism:

As broad as the congressional enforcement power is [under the fifteenth amendment], it is not unlimited. Specifically, ... the power granted to Congress was not intended to strip the States of their power to govern themselves or to convert our national government of enumerated powers into a central government of unrestrained authority over every inch of the whole Nation.

Oregon v. Mitchell, 400 U.S. 112, 128, 91 S.Ct. 260, 266, 27 L.Ed.2d 272 (1970).

The State of Texas has chosen to have trial judges who wield full judicial authority alone, a structure we must accept. Subdistricting would not create an equal opportunity for representation in decision-making, for

[t]here can be no equal opportunity for representation within an office filled by one person. Whereas, in an election to a multi-member body, a minority class has an opportunity to secure a share of representation equal to that of other classes by electing its members from districts in which it is dominant, there is no such thing as a "share" of a single-member office.

Butts, 779 F.2d at 148. What subdistricting does, rather than provide minorities with representation in all decisions, is to simply allocate judges, and thus judicial decisions, among various population groups. The Voting Rights Act does not authorize such allocation. It cannot be made to authorize allocating judges by simply restating the office of district judge as a shared office or by asserting that the "function" of an office is not relevant. Saying that district judges in fact share a common office that can be subdistricted does not make it so. Nor does the assertion that function is not relevant make sense. Function is relevant to the threshold question of what features of the state arrangement define the office.

These judges all hear and decide their own docket of cases, and their character as single-office holders instead of members of a multi-member body is emphasized by the problems inherent in attempting to break the linkage of jurisdiction and elective base. To do so may well lessen minority influence instead of increase it, surely not what Congress intended when it enacted the Voting Rights Act or its amendments. The current system of electing district judges at least permits voters to vote for each and every judicial position within a given district, generally a county. It is more likely, therefore, that minority voters will have some influence on the election of each judge. Under the district court's order, each voter would have the opportunity to vote for only one judge in each district, the judge whose position was assigned to the subdivision. At the same time, a minority litigant will be assigned at random to appear before any district judge in the county. Under the district court's orders it is much more likely than not that a minority litigant will be assigned to appear before a judge who is not elected from a voting district with greater than 50% minority population. Instead, the great majority of district judges will be elected from new voting subdistricts with negligible minority populations and, consequently, negligible minority political influence on the outcome of those elections. Under the new order requiring election of judges from subdistricts, 9 of the 59 judicial positions in Harris county would be elected from minority-dominated subdivisions. Minority voters would have very little influence over the election of the other 50 judges, for the minority population is concentrated in those 9 subdivisions. When minority members are litigants, however, they would not necessarily appeal before one of the judges elected from a minority-dominated subdivision. Instead, a minority member would have an 84.75% chance of appearing before a judge who has little direct political interest in being responsive to minority concerns. The minority member would have a 98.3% chance of appearing before a judge in whose election he had not been able to vote. This is not like the situation in Chisom, where the judges were all part of one body, and every case that went to the Louisiana Supreme Court was heard by all of the judges, so every individual litigant from the state of Louisiana was assured that a judge for whom he had an opportunity to vote would hear his case.

Requiring subdistricting for purposes of electing district judges, unlike other offices, would change the structure of the government because it would change the nature of the decision-making body and diminish the appearance if not fact of its judicial independence — a core element of a judicial office. Trial judges would still exercise their full authority alone, but that authority would no longer come from the entire electorate within their jurisdictional area. Subdistricting would result in decisions being made for the county as a whole by judges representing only a small fraction of the electorate. This does not occur when members of larger bodies are elected from subdistricts, for when the body makes a decision, the interests of all electors are still represented in each decision. When the decisions are not made by a group, the nature of the decision-making body as representative of all of the electors is fundamentally changed through subdistricting. The State of Texas has struck for the essential and defining quality of independence by defining the office of trial judge as a person who judges singly and whose power is derived from an electoral base equal to jurisdictional base. Trial judges are not members of a multi-member body, although there are many district judges, for the district judges do not decide cases as a body. Disregarding the state's insisted linkage of elective base and jurisdiction for single office holders by subdistricting or ignoring their discrete activity, causes a fundamental change in the very office of district judge, a result not contemplated by the Voting Rights Act. These elements define the office; they are far more than the "manner" of election.

One can view the single-official doctrine as being no more than a statement of the mechanical impossibility of gaining greater representation for minorities. This approach is simply a resignation to the reality that if there is only one official, there can only be an at-large election. A second view is that the single official exception expresses far more. This view recognizes that we must accept the state's definition of the office, and that where functions are singly exercised, providing single-member districts is no more than proportional representation in its most superficial form.

Some district courts have proceeded with the first view, concluding that the single official doctrine is inapplicable where more than one official was elected at-large by the same electorate. It is plain that this entire suit rests upon the premise that the single official exception reflects no more than the reality that there is nothing to divide unless there is more than one judge in a single county. It is no accident that this suit attacks only the nine counties with multiple district judges and minority populations. But, the right secured to minorities under Section 2 of the Voting Rights Act to not have their vote diluted is expressed in the assertion that their interests are to be represented in governmental decisions. Where judges make their decisions alone, electing judges from single member districts only increases the likelihood that a small number of governmental decisions will be influenced by minority interests, while minority interests will not be represented at all in the majority of judicial decisions. In this way subdistricting would work a fundamental change in the scheme of self governance chosen by the state of Texas, for it would change the authority behind the decisionmaking body of the Texas courts — and in doing so it would retard, not advance the goals of the Voting Rights Act.

In sum, the single-official concept as we apply it here, whatever its full import in other contexts, is no more than a specific application of the basic principle that analysis under the Voting Rights Act proceeds without changing the state's definition of the office. With the judges acting alone, each judge the decision-making body, a coterminous electoral and jurisdictional base is a core component of the office. Subdistricting would change that office in ways wholly different from changing the selection of members of a governing body as distinguished from the body itself.

D.

Plaintiffs argue that the state's interest in linking jurisdiction and elective base is weakened because in 1985 Texas granted authority to counties to provide for the election of district judges from smaller geographical units. There are two difficulties with this argument. First, no county has elected to do so, and, second, the change only allows the creation of districts smaller than a county. It does not purport to authorize the election of district judges with countywide jurisdiction from districts smaller than the county.

It is also suggested that there is no unacceptable appearance of bias (translate, you still have a court of law) in the prosecution of claims where one litigant is a constituent of a district judge and the other is not. The argument continues that such a circumstance is presented where one of the parties is from another county. This suggestion ignores the fact that the state recognized that elimination of this risk and appearance of bias was essential to the office it was creating by an elaborate set of rules controlling venue. Indeed, Texas has perhaps the most developed venue practice of any of the states, doubtlessly attributable to its diversity and size, allowing a mini-trial of venue facts. Whether a trial proceeds in the plaintiff's home county in El Paso or a defendant's home county in Dallas is of great moment. In sum, the intercounty bias argument proves, rather than defeats, the point. Avoiding the fact and appearance of bias is a powerful state interest. There is no corresponding system of venue rules for a subdistricted county. Rather, as we explained, the state insists on linking the elective and jurisdictional base. Texas wants a trial judge, not a partisan. We are persuaded that Texas has a compelling interest in linking jurisdiction and elective base for judges acting alone. By definition there can be no dilution from the county-wide election of such single officials.