dissenting.
The majority concludes that the “single-subject” rule is not violated by this initiative’s two topics, (1) the repeal of state statutes regulating transportation and (2) the directive to the Governor to “use best efforts” to persuade the U.S. Congress to repeal the federal Jones Act. Additionally, the majority concludes that the Jones Act provision of the initiative enacts a “law” and is thus a proper subject for an initiative under Article XI, § 1 of the Alaska Constitution. I dissent from both of these conclusions.
First, this initiative violates the purposes of the single-subject rule. This court has mistakenly continued to give the rule such an extremely liberal interpretation that the rule has become a farce. Prior decisions tell us that in Alaska any hodgepodge of laws will comply if broadly construed to “relate to” such general topics as water, or land, or taxation. Having made the rule almost meaningless as applied to legislative enactments, this court has extended the same liberal interpretation to the initiative process. In cases to come, what contrived combination of proposed laws, covered by a broad title, would not comply with the majority’s view of the single-subject “rule”? 1
*1183Second, the general language of Article XI, § 1 of the Alaska Constitution clearly limits the initiative process to enacting “laws,” which in turn are further restricted by the specific provisions of Article XI, § 7. Part of the proposed initiative presents a mere resolution that does not qualify as a proposed law and should at least be severed from the rest of the initiative. The drafters of the initiative apparently foresaw this possibility and specifically provided that these independent provisions were severable. At the very least, this initiative’s two distinct provisions should be severed and presented on the ballot as two independent proposals, one to effectively eliminate certain state transportation regulations, and the other to express a desire for federal repeal of the Jones Act.
I.
Turning first to the majority’s main conclusion, I cannot agree that this initiative’s two topics can be said to “fairly relate” to the same subject.2
In Suber v. Alaska State Bond Committee, 414 P.2d 546 (Alaska 1966), this court stated that:
The purpose of the constitutional requirement that every bill be confined to one subject ... is to prevent the inclusion of incongruous and unrelated matters in the same bill in order to get support for it which the several subjects might not separately command, and to guard against inadvertence, stealth and fraud in legislation.
Id. at 557 (emphasis added).
As this court explained later, the primary aim of “one-subject” provisions in state constitutions is to restrain “log-rolling” in the legislative process. Gellert v. State, 522 P.2d 1120, 1122 (Alaska 1974).
Since its earlier decisions, this court has consistently been extremely liberal in finding that various combinations of statutory provisions could be construed to “fairly relate” to the same subject.3 For example, the Gellert court added that the constitutional provision should be construed with “considerable breadth.” Id. This court has since construed it with such “breadth” as to render the rule meaningless. Moreover, this majority opinion pays little heed to the second purpose of the single-subject rule, i.e., “to guard against inadvertence, stealth and fraud.” This decision invites stealth or fraud by initiative promoters, and inadvertence by voters, by stretching the rule to allow contrived pairings of a serious proposal to repeal a state law with a vague directive to the Governor to lobby Congress to repeal a locally unpopular federal act.
The majority’s view of the single-subject rule actually allows disparate subjects to be enfolded within the cloak of a broad generality.4 In Gellert, for instance, this *1184court decided that two topics, a flood control project in Fairbanks and small boat harbors for coastal villages, both pertained to “one ongoing plan” for water resources development. As Justice Fitzgerald pointed out in his dissent, while it was true that all of the projects included in the proposition involved water, so do dams, bridges, sewer systems, and many other things. Justice Fitzgerald explained that:
It is admittedly difficult to determine whether a group of projects has one subject matter but difficulty of enforcement does not discharge the duty of this court to ensure that constitutional limitations are observed. And although the judgment of the legislature is entitled to great weight on the issue, it is finally for this court to make the determination. As is accurately noted by the majority, the purpose of the one-subject requirement is to prevent logrolling, i.e., the assembling of a number of pet projects into one bill to consolidate the support for each to achieve a sufficient total. The proposition in question is" a good example of what the constitutional provision sought to avoid. It is designed to gather voter support for a project in the interior of Alaska by linking it with harbor projects dear to the coastal towns and villages. This is justified, according to the majority, because federal funding is available and the projects are part of a Corps of Engineers program for the state of Alaska. I do not agree with the logic of the majority that such considerations justify the interpretation which is now to be applied, since for all practical purposes it renders the constitutional provision meaningless.
Id. at 1124.
In this court the liberal interpretation of the single-subject rule was first developed in cases dealing with legislative enactments, such as Suber v. Alaska State Bond Committee in 1966 and Gellert v. State in 1974. In cases after Gellert, this court continued to accept strained combinations of subjects as complying with the rule.5 After the single-subject rule was applied to initiatives, this court continued to apply the same liberal interpretation of the rule. This court has since decided to apply the same extremely deferential standard of review to initiatives that it applies to legislation. However, the two processes are quite different. Ideally, if the purposes underlying the single-subject rule are to be realized, courts should consider the contexts to which this constitutional provision applies.6
Whenever a bill becomes law through the initiative process, all of the problems that the single-subject rule was enacted to prevent are exacerbated. There is a greater danger of logrolling, or the deliberate intermingling of issues to increase the likelihood of an initiative’s passage, and there is a greater opportunity for “inadvertence, stealth and fraud” in the enactment-by-initiative process. The drafters of an initiative operate independently of any structured or supervised process. They often emphasize particular provisions of their proposition, while remaining silent on other (more complex or less appealing) provisions, when communicating to the public. *1185Indeed, initiative promoters typically use simplistic advertising to present their initiative to potential petition-signers and eventual voters. Many voters will never read the full text of the initiative before the election. More importantly, there is no process for amending or splitting the several provisions in an initiative proposal. These difficulties clearly distinguish the initiative from the legislative process.
To the extent that it is rational at all, this court’s continuing refusal to rigorously apply the single-subject rule to legislation represents an understandable deference to the legislature as a concordant body of government. It also represents a respect for the legislative process itself: the elaborate procedure by which a bill originates, is reviewed by legislators and experts, and ultimately becomes law. There are no such safeguards, no such review process, between the filing of an initiative petition and its submission to the electorate.7 A procedural vacuum does not deserve great deference.
I must emphasize that my position here suggests no disrespect of the people’s right to make decisions and act in their own sovereign capacity rather than through their elected representatives. Indeed, the opposite is true. By seriously implementing the single-subject rule, this court would mandate that the essence of the initiative process be respected. We would insure that the will of the people is accurately and effectively expressed. In contrast, the majority’s opinion permits initiative-drafters to juxtapose disparate issues in a single petition. This court thus reduces the likelihood that the eventual vote on the initiative will precisely reflect the will of the Alaska electorate.
In this case, the majority strays even farther from the constitutional provision by finding that the single-subject rule is somehow not violated by this initiative’s pairing of a law and a non-law, the pairing of (1) the proposed repeal of state statutes regulating transportation and (2) the anti-Jones-Aet directive to the Governor.
Our so-called standard in deciding whether a proposed law complies with the single-subject rule has been repeatedly recited by this court:
Ultimately the decision in cases of this kind must be made on a basis of practicality and reasonableness. In determining whether a bill is confined to one subject, we agree with the statement:
All that is necessary is that the act should embrace one general subject; and by this is meant, merely, that all matters treated of should fall under some one general idea, be so connected with or related to each other, either logically or in popular understanding, as to be parts of, or germane to, one general subject.
Short, 600 P.2d at 24 (quoting Gellert, 522 P.2d at 1123, quoting Johnson v. Harrison, 47 Minn. 575, 50 N.W. 923, 924 (1891)).
First of all, this standard seems to be no standard at all. It says that all matters falling “under some one general idea” must be connected or related, either logically or in popular understanding.
Basically, the Gellert “test” would be more useful if stripped of its overmodify-ing circumlocution. A stronger and clearer version of the Gellert “test” would read as follows:
An act or initiative should embrace one subject.- By this we mean that all matters treated should be logically connected.
Naturally, merely rewording the existing “test” does not automatically turn this court away from the Anything Goes approach of the “merely ... germane” stan*1186dard embraced in Gellert. But it is a necessary start. Then we must interpret what “connected” means. Certainly provisions that are “inextricably intertwined” are connected, and it is reasonable to infer from the intertwining language in North Slope Borough v. Sohio, 585 P.2d at 545, that this court was focusing on whether provisions have an impact on one another. I favor requiring provisions to be reasonably interdependent. Still, no articulation of a revised standard will do away with debates as to what amounts to connectedness, or intertwining or interdependence. Nevertheless, legislators and initiative promoters should realize that enactments should be presented clearly and candidly. Toward that end, we need to use a plainer standard and be more willing to look closely at the logic of an asserted connection and the reasonable interdependence of separate provisions. Our ruling should be that the court seriously means to effect the purposes of our constitution’s single-subject rule, to discourage logrolling and the type of duplicity evident in the anti-A.T.C. initiative before us.
Still, even under the vague tautology that we adopted as our limited standard of review in Gellert, I cannot agree that the two parts of the initiative at issue here can be said to fairly relate to the same subject, either “logically” or in “popular” understanding. What would determine the logic of the asserted connection between the two parts, the abolition of the Alaska Transportation Commission and the expression of anti-Jones-Act wishes to the Governor?
The majority states that it does not matter that the repeal of the state laws does “not interact or interrelate legally” with the repeal (actually the expressed wish for the repeal) of the federal law. They contend that the alleged interaction is self-evident, adopting the “perspective of the miner at Minto who wants to bring his supplies from Seattle.” Is this logical? Any combination of things, ideas, laws and resolutions can be labelled as a “subject” as long as someone so classifies them. The majority notes that the two sources of regulation of transportation, the Alaska Transportation Commission and the federal Jones Act, “may well be perceived by deregulators as one of this state’s most serious problems.” The perception, or labelling, advanced by the proponents of an initiative should not determine whether its different parts relate fairly and logically to a single subject.8 The majority’s treatment of this matter focuses not on the initiative’s subject {s) but rather on its general policy objectives and its drafters’ perceptions. The initiative addresses two different subjects that its proponents seek, by a broad title or policy statement, to place under a single umbrella. But the single-subject rule is concerned with subjects, not with catch-all titles.
Harking back to Gellert, the majority states that the two different parts of this initiative are, from one viewpoint, “inextricably related, certainly far more integrated and related than boat harbors and flood control projects.” I disagree. At least the legislation at issue in Gellert did not misleadingly mix the enactment of a state law with a mere directive reflecting popular sentiment of the repeal of an unpopular federal act.
In response to the concern that we not apply a standard to the initiative process *1187different from that which we apply to acts of the legislature, this court should take this opportunity to set forth a stronger standard, prospectively applicable to both forms of legislation. CFEC v. Byayuk, 684 P.2d 114, 117-118 (Alaska 1984); Division of Elections v. Johnstone, 669 P.2d 537, 542-544 (Alaska 1983); Warwick v. State, 548 P.2d 384, 393 (Alaska 1976). Although I agree with Justice Burke’s willingness to apply retroactively the ruling we favor, an acceptable argument can be made to the effect that the anti-ATC/anti-Jones-Act promoters were relying on our prior decisions in this area, such that they had reason to believe that Anything Goes, even linking together a law to abolish the state’s transportation commission and a plebiscite on popular sentiment against a federal law affecting interstate shipping. A prospective application would be appropriate here.
II.
As for the majority’s secondary conclusion, I cannot agree that the initiative’s “Jones Act” provision can plausibly be construed to create a “law.” The majority observes that “nothing in the constitution of this state limits the legislature’s power to enact laws establishing as this state’s policy a change in existing federal law,” but this observation begs the question: What distinguishes a “law” from a resolution?
Article XI § 1 of the state constitution clearly states that the people may propose and enact “laws.” Simply because an initiative provision tells the Governor to try “to persuade” Congress to repeal a federal law does not mean that provision establishes anything more than an expression of policy.
The majority mistakenly compares this initiative provision to the Territorial Legislature’s enactment of a law that established the Alaska Statehood Committee to lobby for statehood for Alaska. But the Jones Act provision of the initiative establishes no such committee, or commission, or board to take specific actions. Instead, this provision presents a plebiscite, a rather vague directive that expresses the popular desire to free Alaska’s commerce from the harmful effects of the federal Jones Act.
The majority also suggests that this plebiscite is law-like because it tells the Governor to do his best to persuade Congress to repeal the Act, since the legislature “frequently enacts laws prescribing the conduct of officials or agencies of state government on matters over which the state has no legal jurisdiction.” But what specific conduct does this initiative provision enforceably prescribe for the Governor? The question again involves inquiry into the meaning of “law.”
Recently, in State ex rel. Brant v. Beer-mann, 217 Neb. 632, 350 N.W.2d 18 (1984), the Nebraska Supreme Court rejected a proposed initiative aimed at expressing the desire of the populace for a nuclear freeze and the forwarding of that expression to leaders in the governments of the United States and the Soviet Union. The proposed initiative was held to be “nothing more than a nonbinding expression of public opinion.” Id. 350 N.W.2d at 23. In Beer-mann the Nebraska court discussed whether a measure seeking an advisory “straw vote” on the electorate’s sentiments on a particular issue is a proper subject for the initiative. The Beermann court reviewed with approval the discussion of a similar question by the Supreme Judicial Court of Massachusetts Opinion of Justices Relative to the Eighteenth Amendment, in 262 Mass. 603, 160 N.E. 439 (1928). In the Massachusetts case the House of Representatives had asked the justices whether a “proposed law” introduced by an initiative petition was really a “law” within the meaning of the initiative provisions of the state constitution. The court held that it was not a law and stated the following:
The word “law” imports a general rule of conduct with appropriate means for its enforcement declared by some authority possessing sovereign power over the subject; it implies command and not entreaty; it is something different in kind from an ineffectual expression *1188of opinion possessing no sanction to compel observance of the views announced. ... The result of the vote as proposed in this initiative petition would be lacking in any effective force_ Superficial appearances cannot clothe with the attributes of law something in substance vain and inoperative. The mandate to the Secretary of the Commonwealth ... to tabulate the returns of the votes and to “transmit copies ... to each senator and representative in congress from this commonwealth” is subsidiary and incidental to the main purpose of the proposed law; it relates to a matter which standing alone possesses no legal force....
160 N.E. at 440 (emphasis added).
In the case before this court, the second sentence of Section 4 of the initiative states that until the Jones Act is repealed, “the governor shall publish an annual report documenting the harmful effects of the Act on Alaska commerce, and progress made towards its repeal.” However, this clearly subsidiary instruction is merely incidental to the main objective of the proposed “law”; it relates to a matter that possesses no legal force, to a directive that merely tells the governor to do his “best” to get the federal Jones Act repealed.
The Alaska Constitution reserves to the people the power to propose and enact “laws” by the initiative process. Our cases have liberally construed this law-making power. Still, even under the most liberal construction, the reserved power of initiative and referendum does not encompass all possible actions of the legislature.9 It does not refer to mere resolutions or expressions of popular wishes. A statute declares law. It must be introduced in the form of a bill, then passed with certain formalities and presented to the Governor for signature. But a resolution of the legislature is just a collective expression of opinion. It certainly is not subject to the same formalities or consequences of enactment.
As recently noted in a pertinent part of a decision of the Supreme Court of California, in A.F.L.-C.I.O. v. Eu, 36 Cal.3d 687, 206 Cal.Rptr. 89, 686 P.2d 609 (1984):
“It is frequently said that the distinction between bills and resolutions is that resolutions are not law. As a generalization this is probably accurate, if by ‘law’ is meant those legislative actions which operate on all persons in society, and must be enforced by the executive department, and sustained by the judiciary.” (1A Sutherland, Statutory Construction (Sands rev. ed. 1972) p. 335.)
Id. 206 Cal.Rptr. at 104, 686 P.2d at 624. I further agree with the California Supreme Court that “a resolution, as distinct from a statute, is essentially an enactment which only declares a public purpose and does not establish means to accomplish that purpose.” Id. 206 Cal.Rptr. at 106 n. 23, 686 P.2d at 626 n. 23.
Although a statute may declare policy, as does a resolution, a statute also provides the means to carry out its substantive provisions. In the case before this court, I cannot agree that the initiative’s Jones Act provision establishes the means to accomplish its purpose. As noted by the California Supreme Court, there is a functional as well as formal distinction between a policy declaration that is part of a “law” or statute and a policy declaration that is simply a resolution. In the former instance, administrative agencies and the courts can cite and rely on the policy declaration when interpreting and applying other statutory provisions. No such legal effectiveness is ascribed to a policy declaration that is simply a conclusory directive to the Governor to do his “best” regarding a subject beyond the state’s legislative jurisdiction. The initiative’s disputed provision amounts not to a “law” but, at most, to a cross between a resolution and a wishful writ of mandamus aimed at the Governor.
Again, our main concern should be that all matters (legislative enactments, initiative petitions and even proposed resolutions) should be presented clearly and honestly to the people of Alaska. As George *1189Orwell observed, a great enemy of clarity is insincerity.10 Here, the binding enactment of an enforceable state law should not be attached to a non-binding provision that implicitly promises to compel the Governor to persuade Congress to repeal a locally unpopular federal act. To do so is simply misleading.
It is probably a good idea to extend the use of the direct vote to resolutions or expressions of policy preferences on matters of statewide, national and even international concern. Indeed, the initiative process seems better suited to express broad declarations of the people’s desires than it is to enact technical, often convoluted, statutes. However, Article XI indicates that the initiative was intended to function as a reserved legislative power, a means of enacting “laws.” Here, the Jones Act provision of the initiative does not meet that standard.
The only way that any vote on such an initiative as this could accurately reflect the will of the people would be to sever the unrelated provisions and present them on the ballot as distinctly independent proposals. This would counter the “piggyback” effect whereby a simplistic resolution regarding a highly unpopular federal law is likely to carry along the enactment of whatever state law is attached for the ride. Any insistence on tying together such independent proposals in one initiative suggests that voters do not deserve a clear opportunity to make decisions on an issue-by-issue basis; it forces voters to accept or reject a package of distinct provisions only on an all-or-nothing basis. Voters deserve better. They deserve to have this court apply a more exacting interpretation of the single-subject rule.
In sum, this initiative violates the original purposes of the single-subject rule and it improperly ties together a law and a mere resolution. To guard against inadvertence by petition-signers and voters and to discourage stealth by initiative drafters and promoters, I would find this initiative invalid.
. Why not allow the linking of car -related laws to (a) require car dealers in Alaska to install airbags in cars sold here; (b) direct the Governor to persuade Congress to make the I.R.S. grant tax breaks to drivers in cold-climate states *1183who buy extra auto-safety options; (c) make vehicular homicide a capital-punishment offense; and (d) direct the Governor to lobby the federal government until it agrees to ban the importation of foreign subcompacts. All of these would-be laws "relate to” cars, and such an assortment of provisions would comply with this court’s extremely liberal interpretation of the single-subject rule.
.Short v. State, 600 P.2d 20, 24 (Alaska 1979). Actually, my difference with the majority might be summed up as turning on different definitions of the word “fairly”; in my view, the various provisions of a single enactment, or initiative, should relate "in a fair manner" to the same subject, not just relate "moderately” or "tolerably" to the same subject. Oxford American Dictionary (1980 ed.); Webster’s Third New International Dictionary (1967 ed.).
. E.g., North Slope Borough v. Sohio Petroleum Corp., 585 P.2d 534, 545 (Alaska 1978); State v. First National Bank of Anchorage, 660 P.2d 406, 415 (Alaska 1982).
. Our application of the single-subject rule to initiatives should also involve a functional test much like that used by the Florida Supreme Court in reviewing recent ballot initiatives proposing constitutional amendments. For instance, in Evans v. Firestone, 457 So.2d 1351 (Fla.1984), the court held that Florida’s single-subject rule (quite like Alaska’s) was violated by a ballot initiative entitled "CITIZEN'S RIGHTS IN CIVIL ACTIONS,” which proposed a state constitutional amendment that would, essentially, encourage summary judgment in cases lacking a genuine dispute as to the material facts and would also limit liability for certain types of damages. The court reiterated its view that “where separate provisions of a proposed *1184amendment are an 'aggregation of dissimilar provisions (designed) to attract support of diverse groups to assume its passage,’ ... the defect is not cured by ... application of an over-broad subject title_” Id. at 1354. Furthermore, part of the Florida court’s test is functional: "where such an initiative performs the functions of different branches of government, it clearly fails the functional test for the single-subject limitation....” Id. In our case the initiative at issue also proposes to perform both executive and legislative functions.
. In First National Bank, this court again recited our so-called standard in deciding whether a legislative act's provisions complied with the single-subject rule. We concluded that in light of prior decisions this court “must likewise conclude that 'land' is not an unduly broad sub-ject_" 660 P.2d at 415.
. As observed in Evans v. Firestone, 457 So.2d 1351, 1357 (Fla.1984), "the initiative process does not provide for a filtering mechanism for the drafting of a proposal through amendments, public debate, and legislative vote. This lack of input in the drafting of an initiative proposal is an important reason for the single-subject limitation.”
. In surveys in other states voters have expressed notable frustration or dissatisfaction regarding the way in which initiative and referendum proposals are presented on the ballot. One study has indicated that more than half of voters (surveyed on their views of the presentation of ballot proposals) reported that only "some of the proposals were clear”. Comment, "The Future of Initiative and Referendum in Missouri," 48 Missouri Law Review 991, 1001, n. 93. Also see D. Butler and A. Ranney, Referendums — A Comparative Study of Practice and Theory (1978).
. It is irrational to focus on the drafters’ stated policy objectives to determine whether an aggregate of provisions complies with the single-subject rule. For example, a plausible sounding policy objective would be "Improvement of the Alaska Economy,” a title under which initiative drafters could present such provisions as (1) setting up a state commission to promote tourism in Alaska; (2) directing the Governor to personally greet each ferry ship bearing out-of-state tourists to the state capital; (3) setting up a state office in Washington, D.C. to promote a constitutional convention for the purpose of amending the equal protection clause or the privileges and immunities clause of the U.S. Constitution, so that Alaska could enact “immigration” requirements to restrict the influx of migrating Outsiders; and (4) raising the fines for drunk driving, with the revenue from such fines to be used to increase the amounts, and decrease the interest rates, of AHFC loans available to Alaska home buyers. From the perspective of such an initiative's promoters, or from the perspective of a tourism-industry employee who wants to buy a home, these incongruous provisions would improve Alaska’s economy.
. Article XI refers only to the enactment of laws and the approval or rejection of legislative acts.
. G. Orwell, "Politics and the English Language,” in 4, In Front of Your Nose (The Collected Essays, Journalism and Letters of George Orwell) pp. 127-140, at 137 (1968).