Is Wikipedia consensus process good?

Wikipedia uses what it calls a method of "consensus". It has two main elements: editing without discussion and discussions, including requests for renaming and requests for comment. Is the process as implemented by Wikipedia good?

Arguments for

 * It is a process that stood the test of time, resulting in an encyclopedia that is hugely popular, covers a huge range of subjects, and often serves as a one-look stop to find some of the best Internet resources on a subject. That is inconclusive yet suggestive.
 * Inconclusive indeed. A different process could work even better.
 * The question is not whether the process is best but rather whether it is good.
 * It is a process that combines elements of voting, the strength of argument and deference to policies. Thus, it eclectically picks some of the best elements from different processes.
 * That is inconclusive. Hybrids are not necessarily best or good. In biology, hybrids often cannot have descendants. (The biology argument does not show the process is bad, it merely warns against the assumption that eclectic hybrids are automatically good.)
 * The idea is not of hybrid, which points to biology, but rather of mixture of elements or to compounds. Pure vote-based process is a single-element one. In chemistry, one can do much more with compounds. But one does not need to invoke any specific analogy from a specific domain; one should simply note the contrast between a single element or atom and a compound object, be it a chemical compound, a composite number, a physical mixture (drinking water) or any other kind of compound.
 * Even so, for some purposes, chemical elements are more useful than compounds, even if they are a physical mixture with trace amounts of other elements or compounds. Metals are a case in point. One should examine specific merits and no dwell on whether the process consists of a single element or multiple elements. The process can turn to be good or better even if predominantly single-element one.
 * Not that it matters so much but steel, a mixture, is much better for many purposes than cast iron. In general, single-element things are more suspect, especially if the suspicion is that the preference for them is driven by a subjective love of simplicity or determinacy and not by objective merits.
 * It is a process that eliminates some of the bad features of democracy by invoking the strength of the argument as a principle. In an ideal world, it is the strength of argument that should decide. A process that does not invoke the strength of argument at all, thus not even hinting at this ideal as an aspirational goal, is inferior.
 * That would mean that Switzerland's direct democracy is inferior to Wikipedia's consensus process.
 * A fair point. However, Wikipedia can afford the exchange of arguments in a discussion; a direct democracy has the discussion before the vote, and then there is the bare vote.
 * And that is what Wikipedia should do as well: have a discussion, and then have a vote.
 * It could do it, and it could work to some extent. However, it would be much more bureaucracy and process delay. Wikipedia combines a discussion with a follow-on implied vote. The follow-on implied vote consists in a particular closure either being contested or not.

Arguments against

 * The name "consensus" as applied to the process violates the common meaning of "consensus".
 * It does. However, the question is not whether the process should be called "consensus" but rather whether it is a good process. It could be called quasi-consensus if desired or other name could be invented.
 * It provides unpredictable rules for closing of requests for comments and similar discussions where the closer has an undue sovereign power to determine the strength of the argument.
 * It may be true to some extent. However, a lone dissenting closure will be quickly reverted. Thus, each closure can be understood as a proposed closure that is then either left unobjected or is objected.
 * Good point. However, some discussions get so few participants that no one notices that the closure was done by a lone voice.
 * When that happens, the issue probably did not matter enough for enough editors.
 * It could, but editor attention is a scarce resource, and editors do not have time to participate in most discussions. The above is inconclusive.
 * The process is dishonestly described as based on the strength of arguments in deference to policy where in fact preponderance of opinion about which arguments are best does play a role.
 * A fair point. However, what it means is that the description of the process is incomplete and confusing, perhaps even deceptive, but it does not mean that the process itself as practiced is bad. What it means that part of that process is de facto left uncodified and that the codification is poor.
 * A fair point. However, what it means in practice is that the result is an indeterminate combination of strength of arguments and preponderance of voices. If the process were based only on vote counts, it would encourage bare voting with little or no attempt at argumentation as often happens e.g. in Wiktionary, where in formal votes many editors are happy to cast their rationale-free votes, and almost no one frowns upon that. A process that explicitly threatens rationale-free votes with being disregarded is a much better process even if it has to pay the cost of indeterminacy and possible quarrels about closures.
 * But what it means in practice is not so much that the strength of the argument wins but rather that the 50% majority wins. Since, it is 50% that reflects the objective power, not any supermajority that is not only not specified but not even hinted to exist.
 * That may be true to some extent. However, an observation of actual requests for comments shows people state and exchange arguments, and link policies applied. Then they discuss which policy has most force and whether it really applies, and supply more evidence. Thus, even if vote count plays a role, the objective of pushing editors to state arguments, provide evidence and apply policies seems to be well met in practice.
 * The process assumes that an unopposed edit has "consensus" in its favor where in fact such an edit is often just unnoticed, and when the opposers did not notice, they could not have opposed.
 * A fair point. However, what it means is again that the process description is bad or incomplete. A perceptive editor will realize the above and will not assume that an unnoticed edit has "consensus" in its favor.
 * Various editors will claim that an edit that was unopposed for a long time therefore has status quo ante. And they are in fact right. That is a problem with the above.
 * A fair point. However, there could be a solution to that problem. The process is not perfect, but it is good overall.
 * In all too many articles decay over time is observed, as unqualified or biased editors tweak articles to fit their preferences, removing valid content or changing neutral wording to a biased one.
 * Fair point. However, it is not clear whether it has to do with the definition of consensus. It may mean there is some potential for improvement, but it is not clear in what direction.
 * Fair point. However, it does not prevent Wikipedia from being a one look for some of the best sources online. Thus, the article is a node to which sources are attached. For this, the article does not need to be accurate. That is a weak argument, though, since the main purpose of Wikipedia is not to serve as one look but rather as an encyclopedia.
 * Biased editors remove useful external links, even those to Britannica.
 * That is not a problem with the consensus process but rather with undercodified rules for external links, and possibly with bad cultural developments concerning external links.
 * The pseudo-consensus process of Wikipedia amounts to 50% majority: that truly reflects the power to close the discussion as having "strong" arguments. One needs to recall that even a minority position can be closed in Wikipedia as supported by "consensus". That means that relatively bad proposals can pass: it is much easier for a plain majority to be wrong than for a 2/3-supermajority to be wrong. By contrast, Wiktionary's 2/3-supermajority for formal votes requires that the proposal is so good and the argumentation is so convincing that it passes the high bar. That is no guarantee against bad proposals since no process can possibly provide such guarantee, but that is some of the best that can be had. In Wiktionary, votes are often preceded by discussions, so there is also the element of discussion: Wiktionary process is not truly single-element one. And in so far as discussion often develops as part of the vote, and in so far as comments of voters influence other voters, Wiktionary process is also a multi-element one, consisting of at least 3 elements.
 * Wiktionary process does not require voters to state rationale, and they often don't. Thus, it does not sufficiently emphasize the need for argumentation and evidence.
 * A fair point. Wiktionary process could be improved by requiring voters to state a rationale or at a minimum to point to a rationale located elsewhere. Still, per above, Wiktionary process creates a much higher bar for bad proposals to pass, and that is a key advantage not to be dismissed.
 * Expanding on the objection above, consensus concerns not only formal votes but also Wiktionary Beer parlour discussions. There, editors find it perfectly acceptable to support or oppose something with almost no rationale. The result is a rule of mindless, thoughtless or inarticulate 2/3-supermajority. Ideally, the mindless or inarticulate editors should be prevented from taking part on the government by the requirement of provision of strong arguments, and by such elements of Wikipedia guidelines that "I don't like it" is not an argument. As a result, Wikipedia process, with all its flaws, does a much better job in letting the articulate and thoughtful govern the project than Wiktionary processes that tend to depend on pain vote counting and have no place codifying that the strength of the argument should play a role.
 * A fair point. However, it assumes that those who are inarticulate or do not wish to produce repetitive material for others to read are in fact thoughtless. They may not be; they could have given enough thought and just do not want to argue about the matter.
 * They may have given proper thought to the matter but they may not have. No evidence of their being able to actually think, reason, examine arguments for and arguments against, perform the necessary taxonomization of the subject, identification of key distinguishing characteristics, etc. is produced. Even a purely repetitive exercise in which someone just reformulates someone else's argument provides evidence of mental and intellectual capability; a plain "support" or "oppose" does nothing of the sort. The requirement of articulation is a very useful filter to remove the most incapable editors (sometimes spelling and grammar correctors) from the government.