Wikipedia:Verifiability/2012 RfC/Closing statements

Procedure
First, a word regarding the process of the RfC itself. The closers have observed that: Consequently all three closers have agreed that the RfC is a valid expression of community consensus.
 * The wording and format was the result of a carefully mediated consensus;
 * The RfC was widely advertised, including two watchlist notices and every important noticeboard; and
 * There was little to no disruption of the discussion process (one sock was found and blocked), and participants were well-behaved.

Numerical results
The following is a summary of the degree of support the individual options and views received.


 * Options:

In the following the results are sorted by descending percentage of support. For this quantitative assessment, qualifiers such as "strong" or "weak" are not taken into account, and neutral opinions do not figure into the percentages indicated.

D. New wording about perceived truth, verifiability. Supported by 51 (62%) – Supported with revisions by 15 (18%) – Opposed by 16 (20%)

C. "Verifiability, not truth" with added clarification. Supported by 15 (31%) – Supported with revisions by 10 (20%) – Opposed by 24 (49%)

B. Recent past version, with "verifiability, not truth". Supported by 20 (43%) – Supported with revisions by 2 (4%) – Opposed by 24 (52%)

A. Current version, with "verifiability, and not truth". Supported by 14 (34%) – Supported with revisions by 4 (10%) – Opposed by 23 (56%)

E. About verifiability, no mention of "truth". Supported by 21 (40%) – Supported with revisions by 0 (0%) – Opposed by 32 (60%)


 * Views:

7. "If the lede includes the phrase 'verifiability, not truth', then I would like it to explain that this means WP's content is determined by previously published information, rather than by the beliefs of its editors." Endorsed by 17 (71%) – Opposed by 7 (29%) – Neutral 7

6. "If the lede includes the phrase 'verifiability, not truth', then I would like it to clarify that this phrase means only that verifiability is a requirement for inclusion." Endorsed by 14 (56%) – Opposed by 11 (44%) – Neutral 6

8. "If the lede includes the phrase 'verifiability, not truth', then I would like it to explain that this does not mean that material must always be included so long as it is verifiable." Endorsed by 14 (56%) – Opposed by 11 (44%) – Neutral 3

5. "I would like the lede to say more than it currently does about the distinction between perceived truth and verifiability." Endorsed by 11 (55%) – Opposed by 9 (45%) – Neutral 6

4. "I don't see any need for the phrase 'verifiability, not truth' to be mentioned on the policy page." Endorsed by 23 (47%) – Opposed by 26 (53%) – Neutral 2

11. "I support a large-scale restructuring/rationalisation/simplification of Wikipedia's policies, such as WP:ATT." Endorsed by 10 (45%) – Opposed by 12 (55%) – Neutral 10

10. "I would like the lede to be just about verifiability, and I don't think it needs to mention 'truth' at all." Endorsed by 14 (44%) – Opposed by 18 (56%) – Neutral 2

12. "I think that it is important that our policies should be immediately and unambiguously clear to all our editors (...)" Endorsed by 15 (41%) – Opposed by 22 (59%) – Neutral 7

3. "I don't think the phrase 'verifiability, not truth' needs to be in the lede itself, but this phrase should be mentioned elsewhere on the policy page." Endorsed by 15 (39%) – Opposed by 23 (61%) – Neutral 6

1. "I think the phrase 'verifiability, not truth' needs to be part of the lede." Endorsed by 21 (37%) – Opposed by 36 (63%) – Neutral 4

9. "I would like the lede to mention 'verifiable but inaccurate' material." Endorsed by 4 (17%) – Opposed by 20 (83%) – Neutral 5

2. "I think the word 'threshold' needs to be part of the lede." Endorsed by 3 (15%) – Opposed by 17 (85%) – Neutral 13

Closing comments by Coren
Making a fundamental change to one of our central policies – especially one of our pillars – is never a light matter. Desire to clean up or streamline them is met with considerable resistance and inertia; and for good cause: regardless of what flaws they might have had, they did work for very many years. This RfC is the result of the long discussion regarding such a change that has been ongoing for years – sometimes with no small amount of acrimony – and I feel it is my duty to interpret it conservatively.

On carefully reading the comments and rationales, there are a number of clear points that stand out:


 * 1) The phrase "verifiability, not truth" holds considerable mind share. It is viewed as a succinct, pithy summary of an ultimately complex concept upon which our encyclopedia its based.  Not unlike a kōan, its apparent contradiction hides a profound insight on the nature of Wikipedia that is sometimes clumsy to make explicit. This can be seen as much in the comments of those who favour retaining the current general form of the lede as it is in many the the comments of those who support a change.
 * 2) It is viewed as critical that such a fundamental policy as WP:V be clear and readily understandable by every editor, and that it cannot be misinterpreted to mean something it does not (whether by genuine confusion or by wilfully twisting it). The lede of the policy is where that critical "first impression" of the policy is formed, and should be exemplary in its clarity.
 * 3) There is currently a clear consensus that "verifiability, not truth" fails at that second point. Even many of those who feel the greatest attachment to the phrase acknowledge that understanding it demands elaboration or explanation to make its point and that – also not unlike a kōan – clear understanding requires insight.

Looking at those three points, then, the distribution of support and opposition to the five lede proposals is a clear consequence and comes at no surprise: there is still a desire and perceived need to retain the valuable "verifiability, not truth" phrase, but its position as the core of the lede is largely seen as confusing or misleading; and needs to be de-emphasised accordingly. Thus, option D has clear consensus for implementation. The matter of how, "verifiability, not truth" is to be referenced there is more divided, but the differences are comparatively minor. The current wording of option D provides an entirely workable compromise that could be further tweaked through the normal editorial process. Some of those points that may need further discussion include whether VnT should be called "historical" or a "slogan", used as a parenthetical in the lede itself, discussed and explained later in the text, or placed in a footnote.

Finally, the second section proposes a number of principles, objectives and views related to the core of the discussion. None of the proposed points have clear consensus in support (except for one that does not apply given the general consensus on the wording of the lede), but the discussion there was illuminating and supportive of the results in the first section.

Therefore, this request for comment should be closed as "Consensus to implement the proposed wording D as the lede to Verifiability". &mdash; Coren (talk) 14:30, 28 July 2012 (UTC)

Closing comments by Sandstein
I first evaluate separately whether any individual options and views have consensus. As to the options, it is apparent at a glance that there is no consensus about options A, B, C and E, which are each supported and opposed by roughly half of the participants. This leaves me to determine whether option D, supported in total by 66 contributors (80%) and opposed by 16 (20%), has consensus. There are a few questionable entries – one entry in the "support with revisions" section is prefaced with "oppose", and two "oppose" opinions are by apparent SPAs, but these cases are so few as to be de minimis. Insofar as the 15 editors (18%) who support option D with revisions suggest any specific revisions, these mostly concern the placement and wording of the footnote mentioning "verifiability, not truth" (VnT). About two people want to remove the footnote, about two want to move it into the lead, and a few others suggest other adjustments; some entries in the "support" section also indicate contradictory preferences related to the VnT phrase. That matter is discussed below with respect to the views concerning that phrase. But on the whole the support for option D is so strong that I find that it enjoys consensus in this RfC.

As to the views, it is apparent that we have consensus for view 7: "If the lede includes the phrase 'verifiability, not truth', then I would like it to explain that this means WP's content is determined by previously published information, rather than by the beliefs of its editors." We also have consensus that the lead should not contain the phrases "verifiable but inaccurate" (per the strong opposition to view 9) or "threshold" (the comments in opposition to view 2 indicate that "requirement" or similar is considered better). Option D is compatible with all of this.

We do not have consensus with respect to the remaining views. Notably, the opinions expressed with respect to the wording regarding VnT indicate that the editorship is split about whether this wording should be part of the lead, either as an expression of the policy's meaning or as a historical reference. This division is most clearly apparent in the discussion about view 1, where 21 editors (37%) believe that it should remain an operative part of the policy (a "mission statement" or a "cornerstone", to quote two opinions), but 36 editors (63%) oppose this, arguing that the phrase is contradictory or can be misread to mean that we don't care whether our content is true. Option D represents a compromise between these two positions in that it moves the phrase into a footnote, using wording that is ambiguous as to whether the phrase remains an operative part of the policy, or whether it is a superseded historical formulation (although I'm not sure whether this makes any practical difference). Consequently, the lack of consensus about the continued significance and use of that phrase does not call into question my finding, above, that we have consensus for option D.

For these reasons, I agree with Coren that this RfC concludes with a consensus to implement option D. The question of whether or how the phrase "verifiability, not truth" should continue to be used to summarize the principle enunciated in the policy's new lead remains unsettled, and therefore it can, if desired, be made the subject of continued consensus-finding processes. But because the compromise wording represented by option D enjoys strong support, any change to it, particularly as regards the contested phrase, should result from a process that is at least as thorough as this RfC, and that concludes with a consensus which is about as clear as the present consensus to implement option D.  Sandstein   17:14, 28 July 2012 (UTC)

Closing comments by jc37
There were many questions presented for discussion in this RfC. Due to the size of the discussion, including the many possible outcomes, and the number of commenters, I went through the discussions, making notes and collating the comments. As I noted here, I intentionally tried to consider what each individual expressed as their thoughts, with the "where" of that of a lesser consideration. (Similar to how "weak" and "strong", while kept in mind, tend to get less of a consideration overall, as they are mostly just indicators of strength of preference.)

There were by my count 128 commenters, and they commented throughout the various discussions at their discretion. And assessing consensus of a single discussion section as if it was the sole factor up for discussion (as if it was a solitary discussion in a vaccuum) would seem to me to be inappropriate, and further would also be counter the the diligent work that the contributors put into making this multi-facted RfC and commenting in it.

With that in mind, I am uncomfortable with the percents listed above on this page, in that in my opinion, they may give the appearance that "vote counting" was/is the determining factor of whether any particular proposal had consensus. (Which obviously is not the case.)

One of the questions in crafting this RfC concerned the belief that there should be no change to the lede - in particular to the phrase "Verifiability not truth" (VnT).

From reading over all the discussion(s), it's fairly clear that "Verifiability not truth" (VnT) has entered the mindset and vernacular (the lexicon?) of the average Wikipedian. Coren explained this rather well in his closing statement (numbers 1, 2, 3).

My assessment of the several discussions in this RfC is that the consensus is that VnT should be kept in some way, but there is no consensus regarding whether VnT should or should not be in the lede, or even in what specific way it should be presented.

There is clearly an openness to discuss the "how" of how VnT should be presented on the policy page. However, there was no consensus on the how of how this should be done. The commenters in the various options had many suggestions for how the wording should be modified. But no particular unmodified option had consensus.

I agree with Sandstein that there was consensus for View #7: "If the lede includes the phrase 'verifiability, not truth', then I would like it to explain that this means WP's content is determined by previously published information, rather than by the beliefs of its editors."

And so, it is not surprising that, related to View #7 having consensus, Option "D" had the most support: in reading the comments of all the commenters, the consensus was again affirmed that VnT - wherever it is placed on the policy page - should be more clearly explained (per view #7). However, there was no real overall consensus for the particular specific "layout" of Option D over any of the others. Indeed, as I noted above, there were many suggestions on how the wording of the various options (including option D) should be modified. However, Option D had the most support over all other proposed options, and so the current wording of Option D provides an entirely workable compromise, as a starting point for determining further consensus concerning the lede (and how to present VnT).

So to recap: There is consensus to retain "verifiability not truth" in some way. There is consensus that it should be more clearly explained. There is no consensus for the precise format of how VnT should be presented. There is consensus that that should be something open to further discussion.

And on that last point, I would suggest a further RfC on that, with similar diligence (and collegiate cooperation) as was used for this RfC.

As a personal aside, though we touch on it in our group statement below, I want to specifically commend all those who so thoughtfully contributed to and participated in this process. - jc37 04:50, 30 July 2012 (UTC)

Result
After examination of the discussion and close consideration of the arguments put forward, closing administrators Coren, jc37 and Sandstein agree that this request for comments concludes with a result of "There is consensus to implement the proposed wording of option D as the lede to Verifiability". The detailed closing arguments and rationales of each closer can be read here.

We note that there remain open questions about the exact status and place of the phrase "Verifiability, not truth", and recommend that the community continue discussion on these points. In particular, jc37 considers that there is consensus for option D only insofar as it forms a basis for such continued community discussion.

We thank each and every participant and organisers of this RfC and the mediation that led to it for their invaluable contributions in this difficult consensus-making exercise.

&mdash; Coren (talk) 03:57, 29 July 2012 (UTC) &mdash;  Sandstein   06:33, 29 July 2012 (UTC) &mdash; jc37 04:53, 30 July 2012 (UTC)